diff --git a/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb b/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb
index 9a622da..b1a13a3 100644
--- a/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb	
+++ b/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb	
@@ -1975,27 +1975,27 @@
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
-       "      <th>63634</th>\n",
+       "      <th>46918</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>108258</th>\n",
+       "      <th>10891</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>82400</th>\n",
+       "      <th>15539</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>229712</th>\n",
+       "      <th>14338</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>239499</th>\n",
+       "      <th>17480</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
@@ -2004,12 +2004,12 @@
        "</div>"
       ],
       "text/plain": [
-       "            Fraud  Normal\n",
-       "63634   578.28934     0.0\n",
-       "108258  578.28934     0.0\n",
-       "82400   578.28934     0.0\n",
-       "229712  578.28934     0.0\n",
-       "239499  578.28934     0.0"
+       "           Fraud  Normal\n",
+       "46918  578.28934     0.0\n",
+       "10891  578.28934     0.0\n",
+       "15539  578.28934     0.0\n",
+       "14338  578.28934     0.0\n",
+       "17480  578.28934     0.0"
       ]
      },
      "execution_count": 30,
@@ -2145,240 +2145,240 @@
        "    </tr>\n",
        "    <tr>\n",
        "      <th>mean</th>\n",
-       "      <td>-0.001558</td>\n",
-       "      <td>0.000875</td>\n",
-       "      <td>-0.001892</td>\n",
-       "      <td>-0.001439</td>\n",
-       "      <td>-0.003383</td>\n",
-       "      <td>0.002990</td>\n",
-       "      <td>-0.002319</td>\n",
-       "      <td>-0.004030</td>\n",
-       "      <td>0.004065</td>\n",
-       "      <td>0.005675</td>\n",
-       "      <td>-0.003996</td>\n",
-       "      <td>0.004484</td>\n",
-       "      <td>-0.001110</td>\n",
-       "      <td>0.002187</td>\n",
-       "      <td>-0.002189</td>\n",
-       "      <td>0.001561</td>\n",
-       "      <td>0.000641</td>\n",
-       "      <td>0.002570</td>\n",
-       "      <td>-0.001781</td>\n",
-       "      <td>-0.005305</td>\n",
-       "      <td>0.003222</td>\n",
-       "      <td>-0.002413</td>\n",
-       "      <td>-0.002843</td>\n",
-       "      <td>-0.004712</td>\n",
-       "      <td>0.004207</td>\n",
-       "      <td>-0.002641</td>\n",
-       "      <td>-0.004198</td>\n",
-       "      <td>-0.001204</td>\n",
-       "      <td>-0.007545</td>\n",
-       "      <td>-0.000781</td>\n",
-       "      <td>0.004354</td>\n",
+       "      <td>-0.001286</td>\n",
+       "      <td>-0.002631</td>\n",
+       "      <td>-0.001593</td>\n",
+       "      <td>0.003897</td>\n",
+       "      <td>0.005106</td>\n",
+       "      <td>0.001136</td>\n",
+       "      <td>0.003338</td>\n",
+       "      <td>0.001420</td>\n",
+       "      <td>0.005280</td>\n",
+       "      <td>-0.003772</td>\n",
+       "      <td>0.001565</td>\n",
+       "      <td>0.006694</td>\n",
+       "      <td>0.006557</td>\n",
+       "      <td>0.000992</td>\n",
+       "      <td>-0.004114</td>\n",
+       "      <td>-0.001910</td>\n",
+       "      <td>0.003052</td>\n",
+       "      <td>0.000119</td>\n",
+       "      <td>0.001319</td>\n",
+       "      <td>0.000586</td>\n",
+       "      <td>-0.001582</td>\n",
+       "      <td>-0.001935</td>\n",
+       "      <td>-0.001926</td>\n",
+       "      <td>-0.001921</td>\n",
+       "      <td>0.000102</td>\n",
+       "      <td>-0.005428</td>\n",
+       "      <td>-0.002525</td>\n",
+       "      <td>-0.001413</td>\n",
+       "      <td>-0.000401</td>\n",
+       "      <td>0.000211</td>\n",
+       "      <td>-0.000611</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>std</th>\n",
-       "      <td>0.998500</td>\n",
-       "      <td>0.991199</td>\n",
-       "      <td>0.990627</td>\n",
-       "      <td>0.992018</td>\n",
-       "      <td>0.997557</td>\n",
-       "      <td>0.990920</td>\n",
-       "      <td>1.000773</td>\n",
-       "      <td>0.970588</td>\n",
-       "      <td>0.952178</td>\n",
-       "      <td>0.997451</td>\n",
-       "      <td>0.994908</td>\n",
-       "      <td>0.996820</td>\n",
-       "      <td>1.001257</td>\n",
-       "      <td>0.997821</td>\n",
-       "      <td>0.997179</td>\n",
-       "      <td>0.998505</td>\n",
-       "      <td>1.001794</td>\n",
-       "      <td>1.018943</td>\n",
-       "      <td>1.004320</td>\n",
-       "      <td>0.993911</td>\n",
-       "      <td>0.942704</td>\n",
-       "      <td>0.939417</td>\n",
-       "      <td>0.989367</td>\n",
-       "      <td>0.965734</td>\n",
-       "      <td>1.002527</td>\n",
-       "      <td>0.997197</td>\n",
-       "      <td>1.002805</td>\n",
-       "      <td>0.986277</td>\n",
-       "      <td>0.933587</td>\n",
-       "      <td>0.954577</td>\n",
-       "      <td>1.046103</td>\n",
+       "      <td>1.000527</td>\n",
+       "      <td>1.006693</td>\n",
+       "      <td>0.997540</td>\n",
+       "      <td>0.997086</td>\n",
+       "      <td>1.006792</td>\n",
+       "      <td>0.991543</td>\n",
+       "      <td>0.996668</td>\n",
+       "      <td>0.973116</td>\n",
+       "      <td>0.948871</td>\n",
+       "      <td>0.993364</td>\n",
+       "      <td>0.990888</td>\n",
+       "      <td>1.002877</td>\n",
+       "      <td>1.000311</td>\n",
+       "      <td>0.997226</td>\n",
+       "      <td>1.005559</td>\n",
+       "      <td>1.002712</td>\n",
+       "      <td>1.001343</td>\n",
+       "      <td>1.003773</td>\n",
+       "      <td>1.005842</td>\n",
+       "      <td>1.000361</td>\n",
+       "      <td>1.006151</td>\n",
+       "      <td>0.966084</td>\n",
+       "      <td>0.995486</td>\n",
+       "      <td>0.993314</td>\n",
+       "      <td>0.999420</td>\n",
+       "      <td>0.995134</td>\n",
+       "      <td>1.001346</td>\n",
+       "      <td>0.979838</td>\n",
+       "      <td>1.026780</td>\n",
+       "      <td>0.956109</td>\n",
+       "      <td>0.993369</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>min</th>\n",
-       "      <td>-1.996495</td>\n",
-       "      <td>-17.189967</td>\n",
-       "      <td>-30.511117</td>\n",
-       "      <td>-20.154062</td>\n",
-       "      <td>-3.927001</td>\n",
-       "      <td>-29.290217</td>\n",
-       "      <td>-15.052954</td>\n",
-       "      <td>-25.218244</td>\n",
-       "      <td>-42.653531</td>\n",
-       "      <td>-8.613050</td>\n",
-       "      <td>-20.376630</td>\n",
-       "      <td>-4.700120</td>\n",
-       "      <td>-18.062021</td>\n",
-       "      <td>-4.027674</td>\n",
-       "      <td>-18.381718</td>\n",
-       "      <td>-4.178028</td>\n",
+       "      <td>-1.996580</td>\n",
+       "      <td>-20.443469</td>\n",
+       "      <td>-23.276581</td>\n",
+       "      <td>-22.213271</td>\n",
+       "      <td>-4.013912</td>\n",
+       "      <td>-23.251009</td>\n",
+       "      <td>-15.288057</td>\n",
+       "      <td>-19.739398</td>\n",
+       "      <td>-33.230188</td>\n",
+       "      <td>-7.955047</td>\n",
+       "      <td>-16.780247</td>\n",
+       "      <td>-4.475685</td>\n",
+       "      <td>-18.698648</td>\n",
+       "      <td>-3.907070</td>\n",
+       "      <td>-20.044245</td>\n",
+       "      <td>-4.397796</td>\n",
        "      <td>-15.182703</td>\n",
-       "      <td>-26.619430</td>\n",
+       "      <td>-29.626400</td>\n",
        "      <td>-11.332636</td>\n",
-       "      <td>-5.576921</td>\n",
-       "      <td>-25.102071</td>\n",
-       "      <td>-30.857650</td>\n",
-       "      <td>-10.220648</td>\n",
-       "      <td>-43.582333</td>\n",
-       "      <td>-4.647752</td>\n",
-       "      <td>-11.098125</td>\n",
-       "      <td>-3.438550</td>\n",
-       "      <td>-21.900531</td>\n",
-       "      <td>-26.225413</td>\n",
+       "      <td>-6.059567</td>\n",
+       "      <td>-30.673398</td>\n",
+       "      <td>-25.367337</td>\n",
+       "      <td>-12.246104</td>\n",
+       "      <td>-37.187338</td>\n",
+       "      <td>-4.636193</td>\n",
+       "      <td>-9.458963</td>\n",
+       "      <td>-3.557744</td>\n",
+       "      <td>-24.515479</td>\n",
+       "      <td>-25.520958</td>\n",
        "      <td>-0.353229</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>25%</th>\n",
-       "      <td>-0.853642</td>\n",
-       "      <td>-0.466209</td>\n",
-       "      <td>-0.362183</td>\n",
-       "      <td>-0.585318</td>\n",
-       "      <td>-0.602321</td>\n",
-       "      <td>-0.501455</td>\n",
-       "      <td>-0.580411</td>\n",
-       "      <td>-0.452014</td>\n",
-       "      <td>-0.174859</td>\n",
-       "      <td>-0.575417</td>\n",
-       "      <td>-0.496465</td>\n",
-       "      <td>-0.742776</td>\n",
-       "      <td>-0.406047</td>\n",
-       "      <td>-0.646881</td>\n",
-       "      <td>-0.446149</td>\n",
-       "      <td>-0.632536</td>\n",
-       "      <td>-0.529489</td>\n",
-       "      <td>-0.569855</td>\n",
-       "      <td>-0.595161</td>\n",
-       "      <td>-0.567157</td>\n",
-       "      <td>-0.272174</td>\n",
-       "      <td>-0.312526</td>\n",
-       "      <td>-0.752276</td>\n",
-       "      <td>-0.260018</td>\n",
-       "      <td>-0.585197</td>\n",
-       "      <td>-0.610906</td>\n",
-       "      <td>-0.683196</td>\n",
-       "      <td>-0.174948</td>\n",
-       "      <td>-0.158387</td>\n",
-       "      <td>-0.330520</td>\n",
+       "      <td>-0.855242</td>\n",
+       "      <td>-0.469524</td>\n",
+       "      <td>-0.358808</td>\n",
+       "      <td>-0.578933</td>\n",
+       "      <td>-0.596381</td>\n",
+       "      <td>-0.500065</td>\n",
+       "      <td>-0.571141</td>\n",
+       "      <td>-0.450554</td>\n",
+       "      <td>-0.172817</td>\n",
+       "      <td>-0.584409</td>\n",
+       "      <td>-0.493729</td>\n",
+       "      <td>-0.741852</td>\n",
+       "      <td>-0.396398</td>\n",
+       "      <td>-0.649772</td>\n",
+       "      <td>-0.449871</td>\n",
+       "      <td>-0.640807</td>\n",
+       "      <td>-0.531302</td>\n",
+       "      <td>-0.575388</td>\n",
+       "      <td>-0.595679</td>\n",
+       "      <td>-0.564300</td>\n",
+       "      <td>-0.272836</td>\n",
+       "      <td>-0.311325</td>\n",
+       "      <td>-0.747634</td>\n",
+       "      <td>-0.262140</td>\n",
+       "      <td>-0.584945</td>\n",
+       "      <td>-0.611699</td>\n",
+       "      <td>-0.681904</td>\n",
+       "      <td>-0.174783</td>\n",
+       "      <td>-0.161692</td>\n",
+       "      <td>-0.331159</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>50%</th>\n",
-       "      <td>-0.215398</td>\n",
-       "      <td>0.010908</td>\n",
-       "      <td>0.039725</td>\n",
-       "      <td>0.114781</td>\n",
-       "      <td>-0.017499</td>\n",
-       "      <td>-0.039237</td>\n",
-       "      <td>-0.207896</td>\n",
-       "      <td>0.031778</td>\n",
-       "      <td>0.018143</td>\n",
-       "      <td>-0.044134</td>\n",
-       "      <td>-0.086174</td>\n",
-       "      <td>-0.028081</td>\n",
-       "      <td>0.139761</td>\n",
-       "      <td>-0.010723</td>\n",
-       "      <td>0.049833</td>\n",
-       "      <td>0.058610</td>\n",
-       "      <td>0.080505</td>\n",
-       "      <td>-0.074360</td>\n",
-       "      <td>-0.007117</td>\n",
-       "      <td>-0.000040</td>\n",
-       "      <td>-0.080443</td>\n",
-       "      <td>-0.041695</td>\n",
-       "      <td>0.003188</td>\n",
-       "      <td>-0.017968</td>\n",
-       "      <td>0.070837</td>\n",
-       "      <td>0.028794</td>\n",
-       "      <td>-0.115727</td>\n",
-       "      <td>0.003776</td>\n",
-       "      <td>0.034129</td>\n",
-       "      <td>-0.264951</td>\n",
+       "      <td>-0.216935</td>\n",
+       "      <td>0.002039</td>\n",
+       "      <td>0.042019</td>\n",
+       "      <td>0.121779</td>\n",
+       "      <td>-0.011481</td>\n",
+       "      <td>-0.041114</td>\n",
+       "      <td>-0.205355</td>\n",
+       "      <td>0.032774</td>\n",
+       "      <td>0.021144</td>\n",
+       "      <td>-0.049309</td>\n",
+       "      <td>-0.082977</td>\n",
+       "      <td>-0.021572</td>\n",
+       "      <td>0.147645</td>\n",
+       "      <td>-0.015403</td>\n",
+       "      <td>0.051120</td>\n",
+       "      <td>0.049099</td>\n",
+       "      <td>0.075530</td>\n",
+       "      <td>-0.077553</td>\n",
+       "      <td>0.000442</td>\n",
+       "      <td>0.006663</td>\n",
+       "      <td>-0.079186</td>\n",
+       "      <td>-0.041587</td>\n",
+       "      <td>0.008966</td>\n",
+       "      <td>-0.018311</td>\n",
+       "      <td>0.067176</td>\n",
+       "      <td>0.021756</td>\n",
+       "      <td>-0.106715</td>\n",
+       "      <td>0.003712</td>\n",
+       "      <td>0.033434</td>\n",
+       "      <td>-0.265271</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>75%</th>\n",
-       "      <td>0.935794</td>\n",
-       "      <td>0.671981</td>\n",
-       "      <td>0.486347</td>\n",
-       "      <td>0.674209</td>\n",
-       "      <td>0.520561</td>\n",
-       "      <td>0.441977</td>\n",
-       "      <td>0.296168</td>\n",
-       "      <td>0.458953</td>\n",
-       "      <td>0.275407</td>\n",
-       "      <td>0.549793</td>\n",
-       "      <td>0.412137</td>\n",
-       "      <td>0.730527</td>\n",
-       "      <td>0.619815</td>\n",
-       "      <td>0.665147</td>\n",
-       "      <td>0.515714</td>\n",
-       "      <td>0.713242</td>\n",
-       "      <td>0.596373</td>\n",
-       "      <td>0.476147</td>\n",
-       "      <td>0.592595</td>\n",
-       "      <td>0.559670</td>\n",
-       "      <td>0.175418</td>\n",
-       "      <td>0.252089</td>\n",
-       "      <td>0.728847</td>\n",
-       "      <td>0.234457</td>\n",
-       "      <td>0.735596</td>\n",
-       "      <td>0.668969</td>\n",
-       "      <td>0.500451</td>\n",
-       "      <td>0.229667</td>\n",
-       "      <td>0.236983</td>\n",
-       "      <td>-0.043378</td>\n",
+       "      <td>0.935268</td>\n",
+       "      <td>0.669606</td>\n",
+       "      <td>0.487402</td>\n",
+       "      <td>0.684229</td>\n",
+       "      <td>0.532341</td>\n",
+       "      <td>0.446344</td>\n",
+       "      <td>0.303910</td>\n",
+       "      <td>0.462273</td>\n",
+       "      <td>0.277804</td>\n",
+       "      <td>0.538129</td>\n",
+       "      <td>0.419252</td>\n",
+       "      <td>0.734912</td>\n",
+       "      <td>0.625491</td>\n",
+       "      <td>0.667426</td>\n",
+       "      <td>0.513534</td>\n",
+       "      <td>0.709356</td>\n",
+       "      <td>0.602099</td>\n",
+       "      <td>0.472357</td>\n",
+       "      <td>0.597576</td>\n",
+       "      <td>0.569193</td>\n",
+       "      <td>0.173913</td>\n",
+       "      <td>0.250213</td>\n",
+       "      <td>0.719569</td>\n",
+       "      <td>0.241384</td>\n",
+       "      <td>0.726921</td>\n",
+       "      <td>0.664707</td>\n",
+       "      <td>0.497974</td>\n",
+       "      <td>0.224292</td>\n",
+       "      <td>0.236219</td>\n",
+       "      <td>-0.045377</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>max</th>\n",
-       "      <td>1.641950</td>\n",
-       "      <td>1.240880</td>\n",
-       "      <td>9.990545</td>\n",
-       "      <td>2.761943</td>\n",
-       "      <td>8.351757</td>\n",
-       "      <td>21.022418</td>\n",
-       "      <td>17.952680</td>\n",
-       "      <td>35.611260</td>\n",
-       "      <td>16.400322</td>\n",
-       "      <td>8.294456</td>\n",
-       "      <td>14.001643</td>\n",
-       "      <td>9.372968</td>\n",
-       "      <td>4.577738</td>\n",
-       "      <td>3.750653</td>\n",
-       "      <td>7.998916</td>\n",
-       "      <td>5.206414</td>\n",
-       "      <td>5.558168</td>\n",
-       "      <td>9.293592</td>\n",
-       "      <td>4.910137</td>\n",
-       "      <td>5.792269</td>\n",
-       "      <td>30.668893</td>\n",
-       "      <td>26.225771</td>\n",
-       "      <td>9.949211</td>\n",
-       "      <td>36.076612</td>\n",
-       "      <td>6.601689</td>\n",
-       "      <td>10.500893</td>\n",
+       "      <td>1.641865</td>\n",
+       "      <td>1.249048</td>\n",
+       "      <td>11.446954</td>\n",
+       "      <td>2.690291</td>\n",
+       "      <td>8.969436</td>\n",
+       "      <td>21.128231</td>\n",
+       "      <td>16.057594</td>\n",
+       "      <td>27.728845</td>\n",
+       "      <td>16.049132</td>\n",
+       "      <td>8.385903</td>\n",
+       "      <td>14.080677</td>\n",
+       "      <td>11.775017</td>\n",
+       "      <td>4.409860</td>\n",
+       "      <td>4.590704</td>\n",
+       "      <td>7.742518</td>\n",
+       "      <td>5.600456</td>\n",
+       "      <td>6.845356</td>\n",
+       "      <td>8.962121</td>\n",
+       "      <td>4.857366</td>\n",
+       "      <td>6.869402</td>\n",
+       "      <td>22.234396</td>\n",
+       "      <td>37.034649</td>\n",
+       "      <td>11.398946</td>\n",
+       "      <td>35.364210</td>\n",
+       "      <td>6.642261</td>\n",
+       "      <td>11.646088</td>\n",
        "      <td>6.475679</td>\n",
-       "      <td>20.039391</td>\n",
-       "      <td>48.865274</td>\n",
-       "      <td>40.424940</td>\n",
+       "      <td>22.795198</td>\n",
+       "      <td>48.080214</td>\n",
+       "      <td>31.149316</td>\n",
        "      <td>21.709793</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
@@ -2388,68 +2388,68 @@
       "text/plain": [
        "               Time            V1            V2            V3            V4  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.001558      0.000875     -0.001892     -0.001439     -0.003383   \n",
-       "std        0.998500      0.991199      0.990627      0.992018      0.997557   \n",
-       "min       -1.996495    -17.189967    -30.511117    -20.154062     -3.927001   \n",
-       "25%       -0.853642     -0.466209     -0.362183     -0.585318     -0.602321   \n",
-       "50%       -0.215398      0.010908      0.039725      0.114781     -0.017499   \n",
-       "75%        0.935794      0.671981      0.486347      0.674209      0.520561   \n",
-       "max        1.641950      1.240880      9.990545      2.761943      8.351757   \n",
+       "mean      -0.001286     -0.002631     -0.001593      0.003897      0.005106   \n",
+       "std        1.000527      1.006693      0.997540      0.997086      1.006792   \n",
+       "min       -1.996580    -20.443469    -23.276581    -22.213271     -4.013912   \n",
+       "25%       -0.855242     -0.469524     -0.358808     -0.578933     -0.596381   \n",
+       "50%       -0.216935      0.002039      0.042019      0.121779     -0.011481   \n",
+       "75%        0.935268      0.669606      0.487402      0.684229      0.532341   \n",
+       "max        1.641865      1.249048     11.446954      2.690291      8.969436   \n",
        "\n",
        "                 V5            V6            V7            V8            V9  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean       0.002990     -0.002319     -0.004030      0.004065      0.005675   \n",
-       "std        0.990920      1.000773      0.970588      0.952178      0.997451   \n",
-       "min      -29.290217    -15.052954    -25.218244    -42.653531     -8.613050   \n",
-       "25%       -0.501455     -0.580411     -0.452014     -0.174859     -0.575417   \n",
-       "50%       -0.039237     -0.207896      0.031778      0.018143     -0.044134   \n",
-       "75%        0.441977      0.296168      0.458953      0.275407      0.549793   \n",
-       "max       21.022418     17.952680     35.611260     16.400322      8.294456   \n",
+       "mean       0.001136      0.003338      0.001420      0.005280     -0.003772   \n",
+       "std        0.991543      0.996668      0.973116      0.948871      0.993364   \n",
+       "min      -23.251009    -15.288057    -19.739398    -33.230188     -7.955047   \n",
+       "25%       -0.500065     -0.571141     -0.450554     -0.172817     -0.584409   \n",
+       "50%       -0.041114     -0.205355      0.032774      0.021144     -0.049309   \n",
+       "75%        0.446344      0.303910      0.462273      0.277804      0.538129   \n",
+       "max       21.128231     16.057594     27.728845     16.049132      8.385903   \n",
        "\n",
        "                V10           V11           V12           V13           V14  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.003996      0.004484     -0.001110      0.002187     -0.002189   \n",
-       "std        0.994908      0.996820      1.001257      0.997821      0.997179   \n",
-       "min      -20.376630     -4.700120    -18.062021     -4.027674    -18.381718   \n",
-       "25%       -0.496465     -0.742776     -0.406047     -0.646881     -0.446149   \n",
-       "50%       -0.086174     -0.028081      0.139761     -0.010723      0.049833   \n",
-       "75%        0.412137      0.730527      0.619815      0.665147      0.515714   \n",
-       "max       14.001643      9.372968      4.577738      3.750653      7.998916   \n",
+       "mean       0.001565      0.006694      0.006557      0.000992     -0.004114   \n",
+       "std        0.990888      1.002877      1.000311      0.997226      1.005559   \n",
+       "min      -16.780247     -4.475685    -18.698648     -3.907070    -20.044245   \n",
+       "25%       -0.493729     -0.741852     -0.396398     -0.649772     -0.449871   \n",
+       "50%       -0.082977     -0.021572      0.147645     -0.015403      0.051120   \n",
+       "75%        0.419252      0.734912      0.625491      0.667426      0.513534   \n",
+       "max       14.080677     11.775017      4.409860      4.590704      7.742518   \n",
        "\n",
        "                V15           V16           V17           V18           V19  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean       0.001561      0.000641      0.002570     -0.001781     -0.005305   \n",
-       "std        0.998505      1.001794      1.018943      1.004320      0.993911   \n",
-       "min       -4.178028    -15.182703    -26.619430    -11.332636     -5.576921   \n",
-       "25%       -0.632536     -0.529489     -0.569855     -0.595161     -0.567157   \n",
-       "50%        0.058610      0.080505     -0.074360     -0.007117     -0.000040   \n",
-       "75%        0.713242      0.596373      0.476147      0.592595      0.559670   \n",
-       "max        5.206414      5.558168      9.293592      4.910137      5.792269   \n",
+       "mean      -0.001910      0.003052      0.000119      0.001319      0.000586   \n",
+       "std        1.002712      1.001343      1.003773      1.005842      1.000361   \n",
+       "min       -4.397796    -15.182703    -29.626400    -11.332636     -6.059567   \n",
+       "25%       -0.640807     -0.531302     -0.575388     -0.595679     -0.564300   \n",
+       "50%        0.049099      0.075530     -0.077553      0.000442      0.006663   \n",
+       "75%        0.709356      0.602099      0.472357      0.597576      0.569193   \n",
+       "max        5.600456      6.845356      8.962121      4.857366      6.869402   \n",
        "\n",
        "                V20           V21           V22           V23           V24  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean       0.003222     -0.002413     -0.002843     -0.004712      0.004207   \n",
-       "std        0.942704      0.939417      0.989367      0.965734      1.002527   \n",
-       "min      -25.102071    -30.857650    -10.220648    -43.582333     -4.647752   \n",
-       "25%       -0.272174     -0.312526     -0.752276     -0.260018     -0.585197   \n",
-       "50%       -0.080443     -0.041695      0.003188     -0.017968      0.070837   \n",
-       "75%        0.175418      0.252089      0.728847      0.234457      0.735596   \n",
-       "max       30.668893     26.225771      9.949211     36.076612      6.601689   \n",
+       "mean      -0.001582     -0.001935     -0.001926     -0.001921      0.000102   \n",
+       "std        1.006151      0.966084      0.995486      0.993314      0.999420   \n",
+       "min      -30.673398    -25.367337    -12.246104    -37.187338     -4.636193   \n",
+       "25%       -0.272836     -0.311325     -0.747634     -0.262140     -0.584945   \n",
+       "50%       -0.079186     -0.041587      0.008966     -0.018311      0.067176   \n",
+       "75%        0.173913      0.250213      0.719569      0.241384      0.726921   \n",
+       "max       22.234396     37.034649     11.398946     35.364210      6.642261   \n",
        "\n",
        "                V25           V26           V27           V28        Amount  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.002641     -0.004198     -0.001204     -0.007545     -0.000781   \n",
-       "std        0.997197      1.002805      0.986277      0.933587      0.954577   \n",
-       "min      -11.098125     -3.438550    -21.900531    -26.225413     -0.353229   \n",
-       "25%       -0.610906     -0.683196     -0.174948     -0.158387     -0.330520   \n",
-       "50%        0.028794     -0.115727      0.003776      0.034129     -0.264951   \n",
-       "75%        0.668969      0.500451      0.229667      0.236983     -0.043378   \n",
-       "max       10.500893      6.475679     20.039391     48.865274     40.424940   \n",
+       "mean      -0.005428     -0.002525     -0.001413     -0.000401      0.000211   \n",
+       "std        0.995134      1.001346      0.979838      1.026780      0.956109   \n",
+       "min       -9.458963     -3.557744    -24.515479    -25.520958     -0.353229   \n",
+       "25%       -0.611699     -0.681904     -0.174783     -0.161692     -0.331159   \n",
+       "50%        0.021756     -0.106715      0.003712      0.033434     -0.265271   \n",
+       "75%        0.664707      0.497974      0.224292      0.236219     -0.045377   \n",
+       "max       11.646088      6.475679     22.795198     48.080214     31.149316   \n",
        "\n",
        "       Amount_max_fraud  \n",
        "count      56961.000000  \n",
-       "mean           0.004354  \n",
-       "std            1.046103  \n",
+       "mean          -0.000611  \n",
+       "std            0.993369  \n",
        "min           -0.046062  \n",
        "25%           -0.046062  \n",
        "50%           -0.046062  \n",
@@ -2656,27 +2656,27 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Epoch: 0 Acc = 0.97776 Cost = 82759.57812 Valid_Acc = 0.97767 Valid_Cost =  8259.38965\n",
-      "Epoch: 5 Acc = 0.98852 Cost = 73093.75781 Valid_Acc = 0.98971 Valid_Cost =  7532.35059\n",
-      "Epoch: 10 Acc = 0.99023 Cost = 62864.16406 Valid_Acc = 0.99094 Valid_Cost =  7449.86719\n",
-      "Epoch: 15 Acc = 0.98994 Cost = 53767.84375 Valid_Acc = 0.99066 Valid_Cost =  7594.30762\n",
-      "Epoch: 20 Acc = 0.98749 Cost = 42582.04688 Valid_Acc = 0.98873 Valid_Cost =  7900.34814\n",
-      "Epoch: 25 Acc = 0.99368 Cost = 36059.70312 Valid_Acc = 0.99442 Valid_Cost =  9603.81641\n",
-      "Epoch: 30 Acc = 0.99562 Cost = 35505.75781 Valid_Acc = 0.99558 Valid_Cost =  13627.77832\n",
-      "Epoch: 35 Acc = 0.99511 Cost = 24553.49219 Valid_Acc = 0.99498 Valid_Cost =  16987.88281\n",
-      "Epoch: 40 Acc = 0.99403 Cost = 15877.84961 Valid_Acc = 0.99438 Valid_Cost =  17617.75195\n",
-      "Epoch: 45 Acc = 0.99708 Cost = 30019.80469 Valid_Acc = 0.99684 Valid_Cost =  21042.43359\n",
-      "Epoch: 50 Acc = 0.99434 Cost = 16480.59766 Valid_Acc = 0.99442 Valid_Cost =  26418.41992\n",
-      "Epoch: 55 Acc = 0.99558 Cost = 10637.00391 Valid_Acc = 0.99589 Valid_Cost =  27701.11328\n",
-      "Epoch: 60 Acc = 0.99739 Cost = 7652.20947 Valid_Acc = 0.99719 Valid_Cost =  35590.22266\n",
-      "Epoch: 65 Acc = 0.99384 Cost = 7286.31445 Valid_Acc = 0.99449 Valid_Cost =  28434.03320\n",
-      "Epoch: 70 Acc = 0.99685 Cost = 5848.18164 Valid_Acc = 0.99687 Valid_Cost =  36722.49219\n",
-      "Epoch: 75 Acc = 0.99669 Cost = 4516.80273 Valid_Acc = 0.99635 Valid_Cost =  39600.40234\n",
-      "Epoch: 80 Acc = 0.99617 Cost = 4747.16602 Valid_Acc = 0.99589 Valid_Cost =  40842.24609\n",
-      "Epoch: 85 Acc = 0.99625 Cost = 4744.38867 Valid_Acc = 0.99617 Valid_Cost =  37780.37500\n",
-      "Epoch: 90 Acc = 0.99820 Cost = 4898.35400 Valid_Acc = 0.99789 Valid_Cost =  49322.94531\n",
-      "Epoch: 95 Acc = 0.99793 Cost = 5422.48340 Valid_Acc = 0.99758 Valid_Cost =  50593.96094\n",
-      "Epoch: 100 Acc = 0.99080 Cost = 7493.91504 Valid_Acc = 0.99020 Valid_Cost =  40995.75781\n",
+      "Epoch: 0 Acc = 0.96045 Cost = 72040.01562 Valid_Acc = 0.96057 Valid_Cost =  5780.09863\n",
+      "Epoch: 5 Acc = 0.97615 Cost = 59136.39062 Valid_Acc = 0.97669 Valid_Cost =  4825.07178\n",
+      "Epoch: 10 Acc = 0.97558 Cost = 49881.53125 Valid_Acc = 0.97640 Valid_Cost =  5390.75293\n",
+      "Epoch: 15 Acc = 0.97695 Cost = 41524.66406 Valid_Acc = 0.97707 Valid_Cost =  6727.43359\n",
+      "Epoch: 20 Acc = 0.97956 Cost = 32224.70703 Valid_Acc = 0.97981 Valid_Cost =  9004.73535\n",
+      "Epoch: 25 Acc = 0.98736 Cost = 24203.04297 Valid_Acc = 0.98739 Valid_Cost =  13070.36719\n",
+      "Epoch: 30 Acc = 0.99214 Cost = 19842.75781 Valid_Acc = 0.99266 Valid_Cost =  18814.29102\n",
+      "Epoch: 35 Acc = 0.99255 Cost = 18297.86133 Valid_Acc = 0.99322 Valid_Cost =  25349.83008\n",
+      "Epoch: 40 Acc = 0.99368 Cost = 10427.76758 Valid_Acc = 0.99389 Valid_Cost =  29455.54688\n",
+      "Epoch: 45 Acc = 0.99122 Cost = 7588.53857 Valid_Acc = 0.99143 Valid_Cost =  32230.46875\n",
+      "Epoch: 50 Acc = 0.99425 Cost = 7408.01416 Valid_Acc = 0.99466 Valid_Cost =  38232.20312\n",
+      "Epoch: 55 Acc = 0.99285 Cost = 6066.94238 Valid_Acc = 0.99294 Valid_Cost =  40489.00000\n",
+      "Epoch: 60 Acc = 0.99652 Cost = 10733.87598 Valid_Acc = 0.99624 Valid_Cost =  46334.16797\n",
+      "Epoch: 65 Acc = 0.99609 Cost = 3727.55029 Valid_Acc = 0.99596 Valid_Cost =  48752.45312\n",
+      "Epoch: 70 Acc = 0.99247 Cost = 5904.94922 Valid_Acc = 0.99256 Valid_Cost =  44731.08594\n",
+      "Epoch: 75 Acc = 0.99488 Cost = 6603.65234 Valid_Acc = 0.99442 Valid_Cost =  50844.75000\n",
+      "Epoch: 80 Acc = 0.99630 Cost = 3647.98730 Valid_Acc = 0.99635 Valid_Cost =  51758.85156\n",
+      "Epoch: 85 Acc = 0.99706 Cost = 3211.94214 Valid_Acc = 0.99702 Valid_Cost =  60267.28906\n",
+      "Epoch: 90 Acc = 0.99712 Cost = 3631.59058 Valid_Acc = 0.99684 Valid_Cost =  62279.23047\n",
+      "Epoch: 95 Acc = 0.99730 Cost = 3395.90649 Valid_Acc = 0.99723 Valid_Cost =  63181.06250\n",
+      "Epoch: 100 Acc = 0.99285 Cost = 5376.84277 Valid_Acc = 0.99277 Valid_Cost =  47035.11719\n",
       "\n",
       "Optimization Finished!\n",
       "\n",
@@ -2769,7 +2769,7 @@
    "outputs": [
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
        "<Figure size 720x288 with 2 Axes>"
       ]
@@ -2818,7 +2818,7 @@
      "output_type": "stream",
      "text": [
       "INFO:tensorflow:Restoring parameters from ./model/model.ckpt\n",
-      "Testing Accuracy = 0.9904498\n"
+      "Testing Accuracy = 0.99227554\n"
      ]
     }
    ],
@@ -4364,13 +4364,13 @@
       "Epoch 5/10\n",
       "227845/227845 [==============================] - 1s 3us/step - loss: 0.0502 - acc: 0.9960\n",
       "Epoch 6/10\n",
-      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0450 - acc: 0.9969\n",
+      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0449 - acc: 0.9969\n",
       "Epoch 7/10\n",
-      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0402 - acc: 0.9972\n",
+      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0401 - acc: 0.9973\n",
       "Epoch 8/10\n",
-      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0377 - acc: 0.9976\n",
+      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0375 - acc: 0.9976\n",
       "Epoch 9/10\n",
-      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0357 - acc: 0.9978\n",
+      "227845/227845 [==============================] - 1s 3us/step - loss: 0.0360 - acc: 0.9978\n",
       "Epoch 10/10\n",
       "227845/227845 [==============================] - 1s 3us/step - loss: 0.0346 - acc: 0.9979\n"
      ]
@@ -4378,7 +4378,7 @@
     {
      "data": {
       "text/plain": [
-       "<keras.callbacks.History at 0x1f5b800d5f8>"
+       "<keras.callbacks.History at 0x14b01327d30>"
       ]
      },
      "execution_count": 59,
@@ -4404,8 +4404,8 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "56962/56962 [==============================] - 2s 38us/step\n",
-      "Test score: 0.027562282392704363\n",
+      "56962/56962 [==============================] - 3s 46us/step\n",
+      "Test score: 0.027564537810652604\n",
       "Test accuracy: 0.9982795547909132\n"
      ]
     }
@@ -4434,7 +4434,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 66,
+   "execution_count": 61,
    "metadata": {
     "_cell_guid": "1d565b39-10cf-48ee-8dad-98f585dc0b23",
     "_uuid": "c59e2bca8f631eb923808c7a7c845beae3265770"
@@ -4446,45 +4446,45 @@
      "text": [
       "Train on 227845 samples, validate on 56962 samples\n",
       "Epoch 1/20\n",
-      " - 1s - loss: 0.0118 - acc: 0.9983 - val_loss: 0.0123 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0331 - acc: 0.9980 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 2/20\n",
-      " - 1s - loss: 0.0114 - acc: 0.9983 - val_loss: 0.0123 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0329 - acc: 0.9980 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 3/20\n",
-      " - 1s - loss: 0.0116 - acc: 0.9983 - val_loss: 0.0121 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0324 - acc: 0.9981 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 4/20\n",
-      " - 1s - loss: 0.0112 - acc: 0.9983 - val_loss: 0.0119 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0317 - acc: 0.9981 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 5/20\n",
-      " - 1s - loss: 0.0115 - acc: 0.9983 - val_loss: 0.0118 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0311 - acc: 0.9981 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 6/20\n",
-      " - 1s - loss: 0.0109 - acc: 0.9983 - val_loss: 0.0118 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0307 - acc: 0.9981 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 7/20\n",
-      " - 1s - loss: 0.0111 - acc: 0.9983 - val_loss: 0.0117 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0306 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 8/20\n",
-      " - 1s - loss: 0.0108 - acc: 0.9983 - val_loss: 0.0118 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0295 - acc: 0.9981 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 9/20\n",
-      " - 1s - loss: 0.0109 - acc: 0.9983 - val_loss: 0.0119 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0296 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 10/20\n",
-      " - 1s - loss: 0.0106 - acc: 0.9983 - val_loss: 0.0118 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0298 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 11/20\n",
-      " - 1s - loss: 0.0111 - acc: 0.9983 - val_loss: 0.0117 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0288 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 12/20\n",
-      " - 1s - loss: 0.0106 - acc: 0.9983 - val_loss: 0.0117 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0292 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 13/20\n",
-      " - 1s - loss: 0.0106 - acc: 0.9983 - val_loss: 0.0116 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0286 - acc: 0.9983 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 14/20\n",
-      " - 1s - loss: 0.0104 - acc: 0.9983 - val_loss: 0.0115 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0288 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 15/20\n",
-      " - 1s - loss: 0.0106 - acc: 0.9983 - val_loss: 0.0117 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0286 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 16/20\n",
-      " - 1s - loss: 0.0106 - acc: 0.9983 - val_loss: 0.0116 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0279 - acc: 0.9982 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 17/20\n",
-      " - 1s - loss: 0.0100 - acc: 0.9983 - val_loss: 0.0115 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0275 - acc: 0.9983 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 18/20\n",
-      " - 1s - loss: 0.0102 - acc: 0.9983 - val_loss: 0.0115 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0272 - acc: 0.9983 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 19/20\n",
-      " - 1s - loss: 0.0101 - acc: 0.9983 - val_loss: 0.0114 - val_acc: 0.9983\n",
+      " - 1s - loss: 0.0270 - acc: 0.9983 - val_loss: 0.0276 - val_acc: 0.9983\n",
       "Epoch 20/20\n",
-      " - 1s - loss: 0.0104 - acc: 0.9983 - val_loss: 0.0114 - val_acc: 0.9983\n"
+      " - 1s - loss: 0.0262 - acc: 0.9983 - val_loss: 0.0275 - val_acc: 0.9983\n"
      ]
     }
    ],
@@ -4495,7 +4495,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 67,
+   "execution_count": 62,
    "metadata": {
     "_cell_guid": "1970b306-ca29-4d69-bef8-78dd0380fc04",
     "_uuid": "10b4e95a8da3549bac6c619f8e3a42a79523d11d"
@@ -4507,7 +4507,7 @@
        "dict_keys(['val_loss', 'val_acc', 'loss', 'acc'])"
       ]
      },
-     "execution_count": 67,
+     "execution_count": 62,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -4519,7 +4519,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 68,
+   "execution_count": 63,
    "metadata": {
     "_cell_guid": "ac30ce3a-89f9-40a4-a392-50f7d8562f9a",
     "_uuid": "0e03766581fb5a6519af77ac9b737869c7b0e28b",
@@ -4528,7 +4528,7 @@
    "outputs": [
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
        "<Figure size 432x288 with 2 Axes>"
       ]
@@ -4568,7 +4568,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 72,
+   "execution_count": 64,
    "metadata": {
     "_cell_guid": "52ed3913-ab3c-48e3-9de7-3810af65283b",
     "_uuid": "0df771cc4e8f0c117aaabd93d9778bec837851d4"
@@ -4606,7 +4606,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 73,
+   "execution_count": 65,
    "metadata": {
     "_cell_guid": "fac6cd1f-36ee-4aaa-af29-f94e69db57ef",
     "_uuid": "3afc3624947e0f5454a3893f86163dea548a70f9"
@@ -4648,22 +4648,18 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "bbde0f17-8207-4182-b515-5272645e489e",
-    "_uuid": "81874f10f4453c3211185dc814351db71087a382"
-   },
+   "metadata": {},
    "source": [
-    "## 2.3 Autoencoder neural network with Keras\n",
+    "## 2.3 Keras Autoencoder 신경망\n",
     "\n",
-    "We have found out that standard neural network is not able to capture this highly skewed Fraud data (< 0.2%).  \n",
-    "So I am going to redo this problem by setting up a **Autoencoder network with Keras **.  By definition, Autoencoder network is such that we **develop a model to predict the input** i.e. we are going to optimize the weights and biases (W and b) so that the model gives f(x) = x. How do we do it in NN? We squeeze the network in the middle with input and output layers being the same. The loss or error in Autoencoder is called reconstruction error which is of course minimized during training.  \n",
+    "저흰 앞의 기본적인 신경망을 통해 심하게 편향된 사기 데이터 (< 0.2%)를 잡을 수 없다는 것을 알았습니다. 그래 이 문제를 소위 인풋을 예측하기위해 만들어진 **Autoencoder 네트워크**를 케라스로 구성하여 이 문제를 다시 바라보기로 하였습니다. 예를 들어, f(x) = x를 모델로 가지기 위해 Weights 와 Biases를 최적화 한다고 해보면, NN에서는 어떻게 할까요? input, output 층이 같아질 때 까지 중간에서 네트워크를 계속 짜낼 것입니다. Autoencoder의 loss와 error는 이러한 학습과정을 최소화 할 수 있도록 재구축 되었습니다.\n",
     "\n",
-    "Let us reload the packages."
+    "패키지들을 재로딩 해보겠습니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 66,
    "metadata": {
     "_cell_guid": "46fd0dfc-0e84-45c2-9693-c07e6ec9fdb2",
     "_uuid": "43c44a4085b5376821650f4e520fddd7c44d10b4"
@@ -4693,38 +4689,33 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 67,
    "metadata": {
     "_cell_guid": "4e21db4e-0eb1-46b0-a72c-72e4492291e0",
-    "_uuid": "0977481fa05b2f9189b4287a74d9f27e74e60af5",
-    "collapsed": true
+    "_uuid": "0977481fa05b2f9189b4287a74d9f27e74e60af5"
    },
    "outputs": [],
    "source": [
     "# Reload data and then drop time column and standardize Amount column\n",
     "# Also we follow the already done analysis on this dataset\n",
-    "df = pd.read_csv(\"../input/creditcard.csv\")\n",
+    "df = pd.read_csv(\"./input/creditcard.csv\")\n",
     "data = df.drop(['Time'], axis=1)\n",
     "data['Amount'] = StandardScaler().fit_transform(data['Amount'].values.reshape(-1, 1))"
    ]
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "3a80bb96-609e-48ac-85ba-327eeaed3a31",
-    "_uuid": "200f4154a82812172bc6750abd17182b0f46c2ea"
-   },
+   "metadata": {},
    "source": [
-    "Training the Autoencoder is somewhat different from what we typically do to. Here we do not specify the split betweeen Fraud and Normal transaction. We train the model on the normal transaction only (see how X_train is selected). "
+    "Autoencoder를 학습하는 것은 기존에 하던 것과는 조금 다릅니다. 일단 여기선 따로 사기(Fraud)인 거래와 보통(Normal) 거래를 따로 구분짓지 않습니다. 보통(Normal) 거래로만 모델을 학습 시키겠습니다. (아래 X_train 코드를 봐보세요!)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 68,
    "metadata": {
     "_cell_guid": "92fea18e-5583-4bb8-b6ba-c7a0c7d464cd",
-    "_uuid": "d38a1b8bb6933057053881c83df26893d3c2354e",
-    "collapsed": true
+    "_uuid": "d38a1b8bb6933057053881c83df26893d3c2354e"
    },
    "outputs": [],
    "source": [
@@ -4739,22 +4730,17 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "13e5058b-7011-4386-be66-a9717d24ed1f",
-    "_uuid": "e352475eb2cf97d07fda9b94df63c74aa6ee0fe1"
-   },
+   "metadata": {},
    "source": [
-    "Below I set up the AUtoencoder model. It has 4 fully connected layers after the input layer with 14, 7, 7 and 29 neurons respectively. First two layers are encoder and the last two layers are decoder. Input and output layers have same number of neurons. We are using L1 regularization.  \n",
-    "We can also build a different network in similar format."
+    "아래에 Autoencoder 모델을 구성하였습니다. 4 FC 층으로 input layer는 14, 7, 7, 29 뉴런으로 구성되어 있습니다. 처음 두 Layer는 Encoder(압축용)이고 마지막 두 Layer는 Decoder(해제용) 입니다. Input과 Output Layer는 같은 수의 뉴런을 가집니다. L1 regularization (Overfitting 방지를 위해) 을 적용합니다. 이런 비슷한 포맷으로 다른 네트워크도 형성할 수 있습니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 69,
    "metadata": {
     "_cell_guid": "a9efb494-f9ac-43ea-97e0-82dbf21bfb30",
-    "_uuid": "4fdf0d3c72e8337887142b87ac1d8b6ebee23644",
-    "collapsed": true
+    "_uuid": "4fdf0d3c72e8337887142b87ac1d8b6ebee23644"
    },
    "outputs": [],
    "source": [
@@ -4771,22 +4757,47 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "185fe5c1-aa05-4e7c-b963-0aa5504a8d57",
-    "_uuid": "b0a2e939953acff569d1cfe233e0aae6ae0334a9"
-   },
+   "metadata": {},
    "source": [
-    "Now we compile the autoencoder model and collect the history for 10 epochs with batch size of 1000."
+    "이 Autoencoder 모델을 Compile하고, batch가 1000인 10 epoch의 기록들을 수집합니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 70,
    "metadata": {
     "_cell_guid": "b6362ffe-6174-4194-b91c-eaea6c1e9905",
     "_uuid": "68953af8780d09095e1cccf95899587ef7fa54fc"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Train on 227451 samples, validate on 56962 samples\n",
+      "Epoch 1/10\n",
+      "227451/227451 [==============================] - 1s 6us/step - loss: 1.5176 - acc: 0.1459 - val_loss: 1.2206 - val_acc: 0.2074\n",
+      "Epoch 2/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 1.0685 - acc: 0.2265 - val_loss: 1.0416 - val_acc: 0.2707\n",
+      "Epoch 3/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.9830 - acc: 0.2880 - val_loss: 0.9903 - val_acc: 0.3025\n",
+      "Epoch 4/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.9440 - acc: 0.3048 - val_loss: 0.9625 - val_acc: 0.3080\n",
+      "Epoch 5/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.9232 - acc: 0.3085 - val_loss: 0.9468 - val_acc: 0.3102\n",
+      "Epoch 6/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.9100 - acc: 0.3119 - val_loss: 0.9367 - val_acc: 0.3134\n",
+      "Epoch 7/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.8998 - acc: 0.3156 - val_loss: 0.9275 - val_acc: 0.3178\n",
+      "Epoch 8/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.8908 - acc: 0.3209 - val_loss: 0.9169 - val_acc: 0.3268\n",
+      "Epoch 9/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.8722 - acc: 0.3474 - val_loss: 0.8963 - val_acc: 0.3654\n",
+      "Epoch 10/10\n",
+      "227451/227451 [==============================] - 1s 5us/step - loss: 0.8588 - acc: 0.3661 - val_loss: 0.8891 - val_acc: 0.3664\n"
+     ]
+    }
+   ],
    "source": [
     "nb_epoch = 10\n",
     "batch_size = 1000\n",
@@ -4800,21 +4811,39 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_uuid": "24b27b01d17d99e3eb3198d3ae51c1df02c12ca0"
-   },
+   "metadata": {},
    "source": [
-    "The model is saved in model.h5 via Keras ModelCheckpoint. We can load the model we want.  \n",
-    "Now let us look at the model loss and accuracy during training and validation."
+    "이 모델은 케라스의 ModelCheckpoint를 통해 model.h5의 형태로 저장되었습니다. 그래서 우리가 원할때 해당 모델을 로딩할 수 있게 되었습니다. 이제 모델의 loss와 accuracy를 training과 validation할 동안 history 된 것을 기반으로 보도록 하겠습니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 71,
    "metadata": {
     "_uuid": "e7eaba597a604fdf84e1c7c53a21ea5a273df289"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<matplotlib.legend.Legend at 0x14b01e29fd0>"
+      ]
+     },
+     "execution_count": 71,
+     "metadata": {},
+     "output_type": "execute_result"
+    },
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1008x576 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "# Plot model loss vs. epoch\n",
     "\n",
@@ -4828,11 +4857,22 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 72,
    "metadata": {
     "_uuid": "40b643c218ef2e087703fe6e489d612b974ebd82"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1008x576 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "# Plot model Accuracy vs. epoch\n",
     "\n",
@@ -4846,22 +4886,107 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_uuid": "94fa7a620b73096f72d3a9040e9685527359d8ad"
-   },
+   "metadata": {},
    "source": [
-    "## Prediction and reconstruction error with Autoencoder\n",
+    "## Autoencoder를 이용한 예측과 error 재구성\n",
     "\n",
-    "Let's do prediction on X_test based on the model and then look at the error (mean squared error) which is called the reconstruction error here."
+    "한번 위의 모델 기반으로 X_test를 예측해보도록 하겠습니다. 그리고 여기서 재구성된 (mean squared) error를 살펴보도록 하겠습니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 73,
    "metadata": {
     "_uuid": "85ef72f49fbb4b1aa012707166266317c9e58f63"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>reconstruction_error</th>\n",
+       "      <th>true_class</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>count</th>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>mean</th>\n",
+       "      <td>0.871004</td>\n",
+       "      <td>0.001720</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>std</th>\n",
+       "      <td>3.568457</td>\n",
+       "      <td>0.041443</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>min</th>\n",
+       "      <td>0.089433</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>25%</th>\n",
+       "      <td>0.330827</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>50%</th>\n",
+       "      <td>0.502962</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>75%</th>\n",
+       "      <td>0.744535</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>max</th>\n",
+       "      <td>254.158931</td>\n",
+       "      <td>1.000000</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "       reconstruction_error    true_class\n",
+       "count          56962.000000  56962.000000\n",
+       "mean               0.871004      0.001720\n",
+       "std                3.568457      0.041443\n",
+       "min                0.089433      0.000000\n",
+       "25%                0.330827      0.000000\n",
+       "50%                0.502962      0.000000\n",
+       "75%                0.744535      0.000000\n",
+       "max              254.158931      1.000000"
+      ]
+     },
+     "execution_count": 73,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
    "source": [
     "predictions = autoencoder.predict(X_test)\n",
     "mse = np.mean(np.power(X_test - predictions, 2), axis=1)\n",
@@ -4871,20 +4996,29 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_uuid": "034da78bc51829cf6a38c6375dbc748bb5267c47"
-   },
+   "metadata": {},
    "source": [
-    "* Although true class is either 0 or 1, reconstruction error has much larger range. Let us look at this reconstruction error distribution for both Normal and Fraud classes in the test dataset."
+    "* 클래스가 0 이든 1 이든, 재구성된 error는 매우 넓은 범위를 가집니다. 이 error의 분포를 보통(Normal) 클래스와 사기(Fraud) 클래스로 나누어 살표 보도록 하겠습니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 74,
    "metadata": {
     "_uuid": "fb31d517e0af7ff89f1f2bbfaf30c805dc5fbee7"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1008x576 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "# Reconstruction error in normal class\n",
     "\n",
@@ -4896,11 +5030,22 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 75,
    "metadata": {
     "_uuid": "f879d4001bc8784fd784c6f17af3af6ca8000b06"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1008x576 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "# Reconstruction error in Fraud class\n",
     "\n",
@@ -4912,11 +5057,18 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_uuid": "ec860edafaea08c950d528e2097c8dd58d50d2cf"
-   },
+   "metadata": {},
    "source": [
-    "We see that reconstruction error has larger tail (worse) in case of Fraud class."
+    "우린 재구성된 에러가 사기(Fraud) 클래스에서 넓은 꼬리를 가진다는 것을 볼 수 있습니다."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Autoencoder의 Confusion Matrix\n",
+    "\n",
+    "거래의 클래스를 예측하기 위해선, 해당 거래의 재구성된 에러를 추정해야합니다. **만약 예측된 에러가 threshold(임계값)보다 커진다면,** 이것은 사기(Fraud)라고 판단하고, 나머지는 보통(Normal) 거래로 판단합니다."
    ]
   },
   {
@@ -4932,10 +5084,9 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 76,
    "metadata": {
-    "_uuid": "c74ffebc84540fe12484e780deaa5974186db1b4",
-    "collapsed": true
+    "_uuid": "c74ffebc84540fe12484e780deaa5974186db1b4"
    },
    "outputs": [],
    "source": [
@@ -4946,11 +5097,22 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 77,
    "metadata": {
     "_uuid": "ff271d2e1ade6d753e6389107b83983d15b2e371"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1008x576 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "# Prediction with a threshold\n",
     "\n",
@@ -4972,11 +5134,22 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 78,
    "metadata": {
     "_uuid": "f7ff959efe78f6f49f44f9b6739db5408bd8ee02"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAs0AAALdCAYAAADXmV4aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvqOYd8AAAIABJREFUeJzs3Xd4VHX2x/HPpBFIwYQSFARD+QkIhAjIUqRLiYDSEQjdCAoqRYTQ2SALCCIGpEhvUlRQYRWU4oKrElCKG4om9IDUhBBInd8fWWYdE7iDJNyMeb+eJ8/jfO+ZO2fQXU+O5/u9FqvVahUAAACAO3IxOwEAAAAgr6NoBgAAAAxQNAMAAAAGKJoBAAAAAxTNAAAAgAGKZgAAAMAARTOQzyUmJmrx4sVq3769atSooerVq6tjx45au3atMjIycvWzrVarpk+frtq1a6t69epatWpVjt5/5MiRevzxx3P0nnlFYmKirly5Yhj33nvv6fHHH9eZM2ceQFYA8NflZnYCAMwTExOjgQMH6uzZs2rTpo06dOig5ORkff311xo3bpz27t2r6dOny2Kx5Mrn79y5Ux988IEaNWqkZs2aqUaNGjl6/y5duqhOnTo5es+84PDhwxo4cKDefvtt1a5d+66xzzzzjEqXLi1/f/8HlB0A/DVRNAP5VHJysl5++WVdu3ZNGzZsUMWKFW3X+vbtq4kTJ2r16tWqVq2aevbsmSs5HD16VJI0dOjQXOkIBwcHKzg4OMfva7Zjx47pt99+cyi2YsWKdn9vAQB/DuMZQD61evVqxcbGatSoUdkWVW+++aYKFy6sDz/8MNdySE1NlSR5eXnl2mcAAJATKJqBfGrz5s0qVKiQnn322Wyve3p6at26ddq4caPdelRUlHr37m3r4vbs2VN79+61i2nSpInGjRunTZs26dlnn1XVqlXVvHlzu5nlJk2aKDIyUpLUtGlTNWnSxLYeGhqaJZ8/rsfHx2vkyJFq1KiRqlSpombNmmnGjBlKTk62xWQ303z27Fm98cYb+tvf/qaqVauqbdu2WrdunV3MyJEj1bJlSx08eFA9evRQUFCQ6tatq4iICN26deuOf6a339u6dWvt27dPXbp0UbVq1dS0aVN98sknSk1N1YwZM1SvXj099dRTev3113X16lW793/xxRfq0aOHatSooSpVqqhJkyaaNm2aUlJSJGXOKI8aNUqS1LNnT9uf2+2cV61apVq1aqlWrVr65ptv7Gaab968aRuD+X2nOioqSpUqVdLQoUPv+t0AID+jaAbyIavVqujoaFWpUkXu7u53jHvsscfk4eFhe/31118rNDRUcXFxGjhwoAYOHKi4uDj17t1bX3/9td17//Wvf2ny5Mlq0aKFRo0apYIFC2rSpEnatWuXJCk8PFzPPPOMJGnUqFEKDw+/p+/w+uuva8eOHerUqZPGjx+vp556SgsWLFBERMQd33P69Gl17NhRX3/9tTp37qwRI0aocOHCGjt2rKZNm2YXe+XKFfXr109ly5bV6NGj9eSTT2rFihWaPXu2YW4XL17UgAEDVKNGDb355ptyc3NTeHi4XnrpJX333Xd6+eWX1bp1a/3zn/+0+9z169frtddek4+Pj4YPH64RI0aoZMmSWrRokRYsWCApc0a5S5cukqQBAwbY/bnFxcVpzpw5GjRokDp37qygoCC7vAoWLKiIiAjduHFD//jHPyRJN2/eVHh4uIoWLapx48YZfjcAyK+YaQbyoatXryotLU3FihVz+D1paWmaNGmSAgIC9NFHH8nb21uS1LVrV7Vu3VoTJ05UgwYNbEV4XFycNm7caBv9eOaZZ/T000/r008/VcOGDdWsWTNFR0dr27ZtatasmUqVKuVwLpcvX9a3336rESNGqF+/fpKkTp06yWq16vTp03d838yZM20z3E888YQkqXv37nr55Ze1ePFitWvXThUqVJCU2ckeM2aMrbvduXNnhYSE6LPPPtOIESPumt+1a9c0duxY9ejRQ5JUqlQphYWF6cSJE/riiy9sv4hER0dr9+7dtvctXrxYwcHBmjt3rm3zZbdu3dS0aVN9+eWXGjRokCpWrKjq1atr7dq1qlu3rt1GwFu3bmn8+PFq3779HXP729/+pi5duujDDz9Up06dtHPnTp08eVILFy7UQw89dNfvBQD5GZ1mIB9yccn8n356errD7/nPf/6j8+fPq3v37raCWZJ8fX3Vo0cPXbhwQYcPH7atBwYG2s1KFytWTEWLFtWlS5fuO38fHx8VKlRIq1ev1pdffqmkpCRJ0pQpU7R06dJs35Oenq6dO3eqfv36toJZyvyzGDBggKxWq7Zv3273nlatWtm9rlixoi5fvuxQjre76FJmx16Snn76abvOfalSpXTx4kXb608//VQLFiywO63k8uXL8vX1tX1HI/Xr1zeMeeONN/TII49o9OjRWrFihbp27aoGDRo4dH8AyK8omoF8qHDhwnJ3d3fonN/bbp/zGxgYmOVa2bJlJUnnzp2zrWV3xJmHh0eOnP3s4eGhSZMm6fLly3r11VdVu3Zt9evXT2vXrrWbaf69q1evKikpKdv8y5UrJylz3vn3/vgdPDw8HP5Fo0iRIra/dnV1zbJ2e91qtdpeu7u76/DhwwoPD1fXrl1Vt25dNWjQQMeOHbOLc/Rz78Tb21tjx47V2bNn5evrqzfffNOhewNAfkbRDORDFotFwcHBOnz4sNLS0u4Y984772jo0KG6ePHiXYu229d+Px99u5udU/5YrLZp00Y7d+7U5MmT1ahRI/30008aN26cOnfubNs0l12O2bldyP++Cyzd33dwc8s6/WZ03vWMGTPUp08fRUdHq1KlSho8eLA2bdqkmjVrOvy5twt0I1FRUZIyf5n440ZOAEBWFM1APvXMM88oKSlJW7Zsyfb6rVu3tGHDBn377bd66KGHVLJkSUmZD0T5o9jYWElSiRIl7jsvFxeXLEVvWlqa3SkTN27cUFRUlCwWizp27Kj33ntP//73v9WzZ08dOXLEbk74Nn9/fxUqVCjX8/+zzp49qwULFui5557TJ598ovHjx+uFF15QxYoVc2Sk5fcOHjyopUuXqmPHjqpQoYLGjRunxMTEHP0MAPiroWgG8qkuXbqoZMmSmjp1qo4dO2Z3LT09XRMmTNClS5f04osvyt3dXU888YSKFSumNWvW2BVYiYmJWr16tYoVK6YqVarcd15FixZVbGys3dFu27dvtxu7OH78uLp3764NGzbY1jw8PFS5cmVJ2XdbXV1d9fTTT2vPnj36+eefbetWq1ULFy6UxWJRo0aN7jv/Pys+Pl6SVL58ebv1Xbt26cSJE3b/ReB2B/zPjLqkpqZq9OjR8vPz05tvvqkJEybowoULmjp16n1kDwB/fZyeAeRTBQoUUGRkpPr27auOHTuqTZs2qlq1qq5du6YvvvhC0dHRatmypfr06SMpc/Ri7Nixev3119WhQwd17NhRkrRhwwb99ttvmj17do6MZLRu3Vp///vf1b9/f7Vt21YnT57UunXrbJ1uSQoKClLNmjX1zjvvKC4uTo8//rji4uK0cuVKlS1b9o6Pzh4+fLi+//57hYaGKjQ0VMWKFdO2bdv03XffqU+fPlkK1gepfPnyeuSRRzRv3jwlJyerRIkSOnjwoD755BMVKFBAN27csMXenrVes2aNLl26pDZt2jj8Oe+//76OHTumGTNmyNfXVzVr1lS7du20bt06hYSE/CUfOw4AOYGiGcjHKleurE2bNmnp0qX65ptvtGXLFlmtVj3++ON666231L59e7s53BYtWmjx4sWaO3eu5syZIzc3NwUFBWny5Mn3NHd7N926dbMdC/f3v/9dFStWVGRkpBYvXmw7QcJisWjOnDmKjIzUjh07tHbtWhUuXFjNmzfXa6+9lmU2+bbSpUtr3bp1mjVrlj788EPdunVL5cqV0+TJk22/BJjFw8NDCxYs0D/+8Q8tX75cVqtVpUuXVnh4uNLS0jR58mQdPnxYVapUUZ06ddSqVSvt2LFD3333nZo3b+7QZxw9elQLFixQ3bp11bp1a9v6G2+8oe3bt2v06NH67LPPeEIjAGTDYnV0SzYAAACQTzHTDAAAABigaAYAAAAMUDQDAAAABiiaAQAAAAMUzQAAAICBv+yRc6mXsj71CwD+qF61PmanAMBJ/HBul9kpSDK/xnEvWtbUzzcLnWYAAADAAEUzAAAAYOAvO54BAADwl5SRbnYG+RKdZgAAAMAARTMAAABggPEMAAAAZ2LNMDuDfIlOMwAAAGCATjMAAIAzyaDTbAY6zQAAAIABimYAAADAAOMZAAAATsTKRkBT0GkGAAAADNBpBgAAcCZsBDQFnWYAAADAAEUzAAAAYIDxDAAAAGfCRkBT0GkGAAAADNBpBgAAcCYZ6WZnkC/RaQYAAAAMUDQDAAAABhjPAAAAcCZsBDQFnWYAAADAAJ1mAAAAZ8ITAU1BpxkAAAAwQNEMAAAAGGA8AwAAwIlY2QhoCjrNAAAAgAGKZgAAAMAA4xkAAADOhNMzTEGnGQAAADBApxkAAMCZsBHQFHSaAQAAAAMUzQAAAIABxjMAAACcSUa62RnkS3SaAQAAAAN0mgEAAJwJGwFNQacZAAAAMEDRDAAAABhgPAMAAMCZ8ERAU9BpBgAAAAzQaQYAAHAmbAQ0BZ1mAAAAwABFMwAAAGCA8QwAAABnwkZAU9BpBgAAAAzQaQYAAHAiVmu62SnkS3SaAQAAAAMUzQAAAIABxjMAAACcCec0m4JOMwAAAGCATjMAAIAz4cg5U9BpBgAAAAxQNAMAAAAGGM8AAABwJmwENAWdZgAAAMAARTMAAABggPEMAAAAZ5LBY7TNQKcZAAAAMECnGQAAwJmwEdAUdJoBAAAAAxTNAAAAgAHGMwAAAJwJj9E2BZ1mAAAAwACdZgAAAGfCRkBT0GkGAAAADFA0AwAAAAYYzwAAAHAmbAQ0BZ1mAAAAwACdZgAAAGdCp9kUdJoBAAAAAxTNAAAAgAHGMwAAAJyI1Zpudgr5Ep1mAAAAwACdZgAAAGfCRkBT0GkGAAAADFA0AwAAAAYYzwAAAHAmVsYzzECnGQAAADBApxkAAMCZsBHQFHSaAQAAAAMUzQAAAIABxjMAAACcCRsBTUGnGQAAADBA0QwAAAAYYDwDAADAmXB6hinoNAMAAAAG6DQDAAA4EzYCmoJOMwAAAGCAohkAAAAwwHgGAACAM2EjoCnoNAMAAAAG6DQDAAA4EzrNpqDTDAAAABigaAYAAAAMMJ4BAADgTDin2RR0mgEAAAADdJoBAACcCRsBTUGnGQAAADBA0QwAAAAYYDwDAADAmbAR0BR0mgEAAAADdJoBAACcCRsBTUGnGQAAADBA0QwAAAAYYDwDAADAmbAR0BR0mgEAAAADdJoBAACcCRsBTUGnGQAAADBA0QwAAAAYYDwDAADAmTCeYQo6zQAAAIABimYAAADAAOMZAAAAzsRqNTuDfIlOMwAAAGCATjMAAIAzYSOgKeg0AwAAAAYomgEAAAADjGcAAAA4E8YzTEGnGQAAADBApxkAAMCZWOk0m4FOMwAAAGCAohkAAAAwQNEMAADgTDIyzP25D4MGDdIzzzxjt7Z792516NBBQUFBatKkiRYvXpzlfYcOHVJoaKiCg4NVv359zZw5U6mpqXYxJ06c0IABA1SzZk3Vrl1b48ePV2Jiol3MpUuXNGzYMNWuXVs1atTQ0KFDdfHiRYdyp2gGAABArtu0aZO2bdtmt7Z//34NGDBAZcuW1Xvvvac2bdpo2rRpWrRokS3m5MmT6t27twoUKKBZs2apb9++WrJkiaZMmWKLiY+PV69evXTp0iVNnTpVw4YN05YtWzRs2DBbTFpamvr166eDBw9qwoQJmjBhgvbv36/+/fsrLS3NMH82AgIAADgTq9XsDO7ZhQsXNHnyZJUoUcJuffbs2apcubKmT58uSWrQoIHS0tI0b948hYaGysPDQwsWLJCPj4/mzp0rDw8PNWzYUJ6enoqIiNBLL72kgIAArVq1SgkJCdq4caP8/PwkSQEBAQoLC9OBAwcUFBSkzZs368iRI9qyZYvKlSsnSapUqZJat26trVu3KiQk5K7fgU4zAAAActWYMWNUr1491alTx7aWnJysqKgoNW/e3C62RYsWSkhI0P79+yVJe/bsUePGjeXh4WGLadmypdLT07V7925bTK1atWwFsyTVr19fXl5e2rVrly2mfPnytoJZku317Zi7oWgGAABArlm/fr1+/vlnjR071m799OnTSk1NVWBgoN16mTJlJEmxsbG6efOm4uLissT4+/vL29tbsbGxkqSYmJgsMa6uripVqtRdYySpdOnStpi7YTwDAADAmZj8RMCEhAQlJCRkWff19ZWvr6/d2tmzZzVlyhRNmTJF/v7+dteuX78uSfL29rZb9/LykiQlJibeMeZ23O2NftevX3copnz58tnGnDx5Mvsv+zsUzQAAAHDYsmXLFBkZmWV90KBBGjx4sO211WpVeHi4GjZsqBYtWmSJt/53NttisWT7OS4uLneNsVqtcnH539BETsXcCUUzAACAMzG509yrVy+1a9cuy/ofu8yrVq3S0aNH9dlnn9lOp7hdBKelpcnHx0eSshwLd/u1j4+PrXv8xxhJSkpKst3D29s725gbN26oZMmShjHZdan/iKIZAAAADstuDCM7X375pa5evar69etnufbEE09owoQJcnV11alTp+yu3X4dGBgoLy8vBQQEZBmfuHz5shITE20zyoGBgVli0tPTdebMGVuXOzAwUMeOHcuSy6lTpxQUFGT4fdgICAAAgBw3ceJEbdiwwe6ncePGKlGihDZs2KCWLVuqZs2a2rp1q60DLWUW2z4+PqpSpYokqV69etqxY4dSUlLsYlxdXfXUU0/ZYr7//ntdu3bNFrN7924lJSWpbt26kjJP0zh+/LhiYmJsMb/88otiYmJsMXdDpxkAAMCZWM0dz3BU2bJls6w99NBD8vDwUNWqVSVJAwcOVJ8+fTRkyBC1a9dOP/74oxYtWqRhw4apYMGCkqT+/ftr8+bNCgsLU69evXTixAnNnDlTnTt31iOPPCJJ6tatm1auXKnevXvrlVde0bVr1zR9+nQ1aNBATz75pCQpJCRE8+bNU//+/TVs2DBZrVbNmDFDFSpUUKtWrQy/j8VqdcITsh2QeinGOAhAvlevWh+zUwDgJH44Z3yW74Nw84Ohpn5+wf4z//R7R44cqX379tk9GXDbtm2aPXu2YmNjFRAQoO7du6tv375274uKitK0adMUHR0tPz8/Pf/88xo8eLDc3d1tMceOHdNbb72lH3/8UV5eXmrWrJlGjBhhN68cFxenyZMna8+ePfLw8FC9evU0cuRIFS9e3DB3imYA+RpFMwBHUTRnup+i2ZkxngEAAOBErBl/yX5nnsdGQAAAAMAAnWYAAABnYvI5zfkVnWYAAADAAEUzAAAAYIDxDAAAAGfiJOc0/9XQaQYAAAAM0GkGAABwJhw5Zwo6zQAAAIABimYAAADAAOMZAAAAzoRzmk1BpxkAAAAwQKcZAADAmdBpNgWdZgAAAMAARTMAAABggPEMAAAAZ2LlnGYz0GkGAAAADNBpBgAAcCZsBDQFRTPytKO/xKpLv1f1Ys8ueqVfD9v6/gOH9e78Zfr5yHH5+nirSYM6eqVfD/k9VNju/T8fOa533l+iA4f/IxcXF9UMrqrhr7yowDKl7OK+3/eTIj9YoaPHY+XtVUjNG9fXq2G9VKhQwXvODYA5Hi5VQpt+WHvXmAEdXtP+f/+kh/wLa/DYgarbpLYKeBbQvm9/1DvjI3XuVJxd/P9VqaBBo19S5aCKSk9P1+6v/q05k+fryqWrdnGO3g+A86JoRp6VlpauMZNnKC0tzW79h/0H9dLQ0fLx9taLPbvIxcVFK9Zt1Pf7DmjlvBkq7OsjSYo9eUZ9Br0pT88CGtCnmyRp2ZqP1fPlYfpo6VwVL1Yk8377DujF10er8uPlNWRgH53/7aJWrtukn48e17I50+XiknWK6U65ATDP1cvXNG5QRJb1Ap4FNDziVV29fE3H//OL3D3c9e6qaXq07KNas2CdblxPUreXOmv+x7PV45l+ir+aIEl6rHwZLdz4ni7GXdLCGUvk5eOlbmGdFVSzikJbvKibSTclyeH7AXBuFM3Isz5YsVa/xJ7Msj7lnffl6uKqlfNmqHSpRyRJTRvWVfueL2vB8g/1xqAXJUkr1n2ipJs3tWzuNFX6v/KSpKdqBOmF/q9r+dpPNHxQf0nS23M+0MMBxbR0zjR5FiggSXo4oLgiZszRnu/36ek6tRzODYB5bt28pS8+3pZlfcjEQXJzd9O4VyJ0PT5Rz3VrrUpBFTWo6zD98E2UJGnP9u+05usl6hbWWe9P/UCSNODNfkpPS1dYu8G2zvKRg0c1a+U0hXRsro+Wb5IkhXRs4dD9gByTwUZAM7AREHnSsV9jNX/ZGg3o3c1u/WzcBR2POaE2LZvYCmZJKlvmUTWqV1uf/vMr29qZc+fl95CvrWCWpKqVHtdDhX11POaEJCk5OUV+DxVWhzYtbQWzJNWsXtWWh6O5Ach7ylUsq85922vzui/00w8HJUnNn2ui07FnbAWuJJ385ZT27t6nZ55rYltLTU3Tlo+22o1i7P/3AUlS+UrlbGuO3g+Ac6NoRp6TOfrwjv5WM1itW9j/C+fCxUuSpAplH8vyvtKlHtbVawmKu3BRklSmVEnFJyTqytVrtpj4hOu6npioYkX8JUkFCnho/swIhfXqanevI8d/lZTZcXY0NwB5z8CR/ZV8K9mu21ux2v/pyKFjWWKPHDquUo+VlE9hb0nS2Jcn6e3R79rF/F+VzF/Cz5+9cM/3A3KMNcPcn3yKohl5zqJV63TqzFmNHzE4y7VCnp6SpBv/nSX8vWvx1yVJl65ckST17d5RAcWKasSEqTr6S6yO/RqrN8b/Q+5u7ureqW22n33u/AVt3LxNU2bNU4Wyj6lpg7oO5wYgbylfqawaNK+nj1d8qsu/Zf7/QsFCBeVT2EcX4y5lib/822VJUomSAVmuFSnur8YhDTQpcox+i7uoT9dsvq/7AXA+zDQjT/kl5qTmLVmt0UNfVonixXQ27oLd9XKBpeXtVUjbdu5R/9DOslgskjLHLPb8sE+SlJKcKkl6uERxvdizi96aOVcder0sSXJ1ddHMiNF2Ixu3xSdcV/MOvSVJBT0LaNSQgSpQwMPh3ADkLR16Pq+0tDStW/Sxba2Qd+aJOLdu3soSn3wrWZLkWcgzy7X1/1opbx8vpaWlacKrb+nq5Wv3dT8AzseUovnChXsrNgIC+C09P0hPT9eYt2bqyWpPqGPbVtnGuLu7q2fX9pq7aKXenDhN/UM7KyM9Q7MXLtfN//5Ly9XNVZL03oLlmr9sjWoGV1Wntq2UnpGhtZ9s1rCxUzRr8mg1qv+3LPefPnGkUtPStGr9Jr34+ihNnzhSzRs/7VBuAPKOAp4eatnhGf1r67d2oxS3T8Ox3uWJatY/bLJydXXV9PBZSk1NVduuIYqYO05FivlrzcL1f+p+wH3jnylTmFI0N2zY0NYhdER0dHQuZoO8Ysnqj3T0lxgtf/9tXb0WL0lKuJ45cnHrVrKuXotXYV8fDej9gq5fT9SqDZ9qy7adkqRG9WqrX49Oeuf9JSrs46OE64lasmaDnqhYQYvenSJX18xCulWzhura7zWNnzpb2556Uh4e/+skF/b1UatmDSVJzRvX1/M9BmjaewvVvPHTDueW3fF0AB68GnWD5eVdSF9/vtNuPSkxSZJUoGCBLO8p4FnALua29PR0/fOjrZKkrz7doYWbIvXSiL7atPrzP3U/AM7JlKL5rbfeuqeiGfnD7u+jlJqaphf6v57l2pLVG7Rk9QZ9uWGpSj4coDdfe0n9Qjvr1OmzKhFQTI+UCNC785fK1dVFj5QormMxsUpJSVXIM41sBbMkubu56dnmjTVz7iLFnDyjihXKZpuLZ4ECalivtlat36Sr1+LvKTcA5qvb9G9KSU7Rnq/+bbd+IzFJCdeuq2jxIlneUzQgc+3ihazzyb+3/fNdCqpVVWXKlVb0waP3fT8AzsGUorl9+/ZmfCzyuDcGvaiE64l2a5evXNXISdPVpmVTtW3ZVEX9/bRl204VLeKvp56spqL+frbYfT8dVuXHK6hAAQ95uGd2kNPTs+7yzfjv40et1gzFnDytAUPHqG/3TuravrVd3I2kJFksFnm4uzucG4C8IahWVf3nwBHdyKbLe+zwcT1e9f+yrD9epYJOxZzR9fhEefl4adk/F2jHll2a89YCu7hC3oUk/W9m2ZH7ATnJymO0TZEnNgIeOnRIe/fuVWpqqm0uLCMjQzdv3lRUVJTWrFljcoZ4EJ6oWCHL2u3NdqUeKaE6tYIlScvXfqJbt5K1Yekcuf13fnnXtz9o/8GfNXnMMElS+cDSKl60iDZt2abuHdvaNvQlJ6fo0y++kt9Dvipf9jFZZFHijSSt3bhZHdq0kLu7u6TMUzS+2rlHNatXlZdXIYdzA2A+VzdXBVYoo42rPs/2+vYtuzTirSGq9XQN7f1X5gbiMuVLq1b9GloWuUqSdOP6DaWmpOjZTi21fM5qW+Hr5eOlti+E6OzJc4o5dsLh+wFwfqYXzWvWrNGkSZOy3UTh4uKi+vXrm5AV8rJ+3TtpyJjJemXEeDVrWFdn437T8rUfq17tGmrdvLGkzI074UMHauiYt9T1xdfUvnULZWRk6JPPtyr25BlNGTtc7m6Z//iPGjJQoyZNV+9XRqh1iya6Fp+gNR99JovFovChA838qgD+hBIlA+RRwMNuA+DvbVq9WZ36tNc/FkzUyvc/1K2byeo+oIt+O39RH36wwRY3LXyW5qybqYWb5mjjys/k7uGudj3aqEgxfw0JffOe7wfkGDYCmsL0onnFihVq0KCBpk2bpvnz5ysxMVHh4eHatWuXRo4cqbZtsz9PF/nXM43ra9rEN7VoxXpNm71ARfz91OeFjurfs7Pd/HKzhvW0cNZkvb9ktWbPXypJqvR/5fX+25NU/281bXGygtxZAAAgAElEQVRtWjSRu5ubFq9ar2nvLVBBT0/9rWZ1vRrWS4+VLvWgvx6A+1TYz1dSZrc4O6kpqXql8xC9Nv4Vhb78gtLTM7T/3z/p3YlzFH81wRb343cH9Fq3N/TisD56eVSY0tPTdeCHQxo9cJKiDxy55/sBcG4W693OyXkAqlatqsjISDVs2FBbtmzRnDlztHlz5qHx77//vnbs2KF169bd831TL8XkdKoA/oLqVetjdgoAnMQP53aZnYIk6cbknqZ+vtfo5aZ+vllMPx/L3d1dnv99yluZMmV08uRJpaZmPpyiRo0aOnHihInZAQAA5DE8RtsUphfNFStW1M6dOyVJgYGBysjI0IEDByTd+0NQAAAAgNxg+kxzr1699Nprr+n69euKiIhQ06ZNNWLECLVq1UqbNm1SjRo1zE4RAAAg72AjoClM7zS3aNFCc+bMUZkyZSRJkyZN0mOPPaZVq1YpMDBQY8eONTlDAAAA5Hemd5olqWnTpmratKkkyc/PT4sXLzY5IwAAAOB/8kTRfPHiRf3nP/9RQkL2R/O0adPmAWcEAACQR/FEQFOYXjRv2bJFo0aNUnJycrbXLRYLRTMAAABMZXrRPGvWLFWpUkXh4eF66KGHzE4HAAAgb2MjoClML5p/++03jR49Wk888YTZqQAAAADZMv30jOrVqys2NtbsNAAAAIA7Mr3TPH78eA0YMECJiYmqVq2aChYsmCWmVq1aJmQGAACQB+Xjp/KZyfSiOSYmRpcuXVJkZKSkzI1/t1mtVlksFkVHR5uVHgAAAGB+0Txt2jSVLVtWL774oooUKWJ2OgAAAHkbGwFNYXrRfP78eU2YMEF16tQxOxUAAAAgW6ZvBKxSpQobAQEAAJCnmd5pHjx4sIYPH66rV6+qWrVq8vLyyhLz5JNPmpAZAABA3mPliYCmML1o7t27tyTpvffek8RGQAAAAOQ9phfNy5cvNzsFAAAA58FGQFOYXjTv2LFDISEhqlq1qtmpAAAAANkyfSPgunXrlJCQYHYaAAAAwB2ZXjQ/8cQT2rNnj9lpAAAAOIcMq7k/+ZTp4xlPPPGEli9frm3btql8+fIqWrSo3XWLxaJJkyaZlB0AAACQB4rmL7/8UsWLF1d6erqOHj2qo0eP2l3//WkaAAAAgBlML5q3b99udgoAAADOw8o5zWYwvWi+LT4+Xj/99JMSExPl7++vqlWrytvb2+y0AAAAgLxRNL///vuaN2+ekpOTbWseHh4KCwvToEGDTMwMAAAgj8nHm/HMZHrRvG7dOs2ePVtdu3ZVmzZtVLRoUf3222/6/PPPNXfuXAUEBKhTp05mpwkAAIB8zPSiedmyZQoNDVV4eLhtrXTp0qpZs6Y8PDy0YsUKimYAAACYyvRzmk+fPq1GjRple61Ro0Y6efLkg00IAAAgD7NmWE39ya9ML5offvhh/frrr9leO378uAoXLvyAMwIAAADsmV40h4SE6N1339W2bdvs1rdu3arIyEi1atXKpMwAAADyIJ4IaArTZ5oHDBigqKgoDR48WB4eHipSpIguX76slJQU1apVS6+//rrZKQIAACCfM71oLlCggFasWKFdu3Zp7969io+Pl6+vr5566ik1aNCAJwICAADAdKYUzZGRkdmue3p6ytPTU5J06NAhHTp0SBaLRa+88sqDTA8AACDvyuCJgGYwpWh+//33DWOsVqus1sy5GYpmAAAAmMmUovnnn3++6/U1a9bo7bffltVq1bBhwx5QVgAAAE4gH2/GM5PpM82/d/r0aY0ZM0Y//PCD6tatq7///e965JFHzE4LAAAA+VyeKZqXLl2qd999V+7u7oqIiFCHDh3MTgkAAACQlAeK5piYGIWHh+unn35SkyZNNGHCBBUvXtzstAAAAPImxjNMYVrRnJGRoYULF2rOnDny8vLSzJkzFRISYlY6AAAAwB2ZUjQfOXJE4eHhio6OVkhIiMaMGSM/Pz8zUgEAAHAqt08Xw4NlStHcsWNHpaeny8fHR1evXr3rCRkWi0WLFi16gNkBAAAA9kwpmoODg21/nZqaakYKAAAAgMNMKZpXrFhhxscCAAA4PzYCmsLF7AQAAACAvI6iGQAAADBg+jnNAAAAuAeMZ5iCTjMAAABggE4zAACAE7HSaTYFnWYAAADAAEUzAAAAYIDxDAAAAGfCeIYp6DQDAAAABug0AwAAOJMMsxPIn+g0AwAAAAYomgEAAAADjGcAAAA4Ec5pNgedZgAAAMAAnWYAAABnQqfZFHSaAQAAAAMUzQAAAIABxjMAAACcCec0m4JOMwAAAGCATjMAAIAT4cg5c9BpBgAAAAxQNAMAAAAGGM8AAABwJmwENAWdZgAAAMAAnWYAAAAnwkZAc9BpBgAAAAxQNAMAAAAGGM8AAABwJmwENAWdZgAAAMAARTMAAABggPEMAAAAJ2JlPMMUdJoBAAAAA3SaAQAAnAmdZlPQaQYAAAAMUDQDAAAABhjPAAAAcCJsBDQHnWYAAADAAJ1mAAAAZ0Kn2RR0mgEAAAADFM0AAACAAcYzAAAAnAgbAc1BpxkAAAC5xmq1aunSpWrRooWqVaumtm3b6rPPPrOL2b17tzp06KCgoCA1adJEixcvznKfQ4cOKTQ0VMHBwapfv75mzpyp1NRUu5gTJ05owIABqlmzpmrXrq3x48crMTHRLubSpUsaNmyYateurRo1amjo0KG6ePGi4feg0wwAAOBEnK3TPH/+fM2ePVuDBw9W9erV9c0332j48OFydXVVSEiI9u/frwEDBqhVq1Z67bXXtG/fPk2bNk1Wq1X9+vWTJJ08eVK9e/dWcHCwZs2apV9//VXvvPOOEhMTNW7cOElSfHy8evXqpWLFimnq1Km6fPmypk+frvPnz2v+/PmSpLS0NPXr109JSUmaMGGC0tLSNGPGDPXv318fffSR3NzuXBpTNAMAACBXpKamavHixXrhhRc0cOBASVKdOnV0+PBhrVy5UiEhIZo9e7YqV66s6dOnS5IaNGigtLQ0zZs3T6GhofLw8NCCBQvk4+OjuXPnysPDQw0bNpSnp6ciIiL00ksvKSAgQKtWrVJCQoI2btwoPz8/SVJAQIDCwsJ04MABBQUFafPmzTpy5Ii2bNmicuXKSZIqVaqk1q1ba+vWrQoJCbnjd2E8AwAAALnC1dVVK1asUFhYmN26u7u7kpOTlZycrKioKDVv3tzueosWLZSQkKD9+/dLkvbs2aPGjRvLw8PDFtOyZUulp6dr9+7dtphatWrZCmZJql+/vry8vLRr1y5bTPny5W0FsyTb69sxd0LRDAAA4ESsGeb+3AsXFxc9/vjjCggIkNVq1aVLl7RgwQJ9++236tKli06fPq3U1FQFBgbava9MmTKSpNjYWN28eVNxcXFZYvz9/eXt7a3Y2FhJUkxMTJYYV1dXlSpV6q4xklS6dGlbzJ0wngEAAACHJSQkKCEhIcu6r6+vfH197/i+rVu36tVXX5UkNWrUSG3btlV0dLQkydvb2y7Wy8tLkpSYmKjr169nG3M77vZGv+vXrzsUU758+WxjTp48ecfcJYpmAAAA52K1mPrxy5YtU2RkZJb1QYMGafDgwXd8X+XKlbVy5UodPXpU7777rsLCwvT6669LkiyW7L+Ti4uLrFbrHWOsVqtcXP43OJFTMdmhaAYAAIDDevXqpXbt2mVZv1uXWZIeffRRPfroo6pVq5a8vb315ptv2griPx4Ld/u1j4+PrXv8xxhJSkpKko+Pj6TMTnR2MTdu3FDJkiUNY7LrUv8eRTMAAAAcZjSG8XvXrl3Tzp07VadOHQUEBNjWK1euLEk6c+aMXF1dderUKbv33X4dGBgoLy8vBQQEZBmfuHz5shITE20zyoGBgVli0tPTdebMGbVo0cIWc+zYsSx5njp1SkFBQXf9LmwEBAAAcCLOtBEwIyNDI0eO1Nq1a+3W9+zZI0mqWrWqatasqa1bt9q6zpL05ZdfysfHR1WqVJEk1atXTzt27FBKSopdjKurq5566ilbzPfff69r167ZYnbv3q2kpCTVrVtXUuZpGsePH1dMTIwt5pdfflFMTIwt5k7oNAMAACBX+Pv7q1u3blqwYIE8PT1VtWpV7du3T/Pnz1enTp1UtmxZDRw4UH369NGQIUPUrl07/fjjj1q0aJGGDRumggULSpL69++vzZs3KywsTL169dKJEyc0c+ZMde7cWY888ogkqVu3blq5cqV69+6tV155RdeuXdP06dPVoEEDPfnkk5KkkJAQzZs3T/3799ewYcNktVo1Y8YMVahQQa1atbrrd7FYf1/W/4WkXooxDgKQ79Wr1sfsFAA4iR/O3f0c3wclrn5jUz//4d077ik+NTVVS5cu1YYNG3Tu3DmVKFFCnTp1Uv/+/W2b77Zt26bZs2crNjZWAQEB6t69u/r27Wt3n6ioKE2bNk3R0dHy8/PT888/r8GDB8vd3d0Wc+zYMb311lv68ccf5eXlpWbNmmnEiBF288pxcXGaPHmy9uzZIw8PD9WrV08jR45U8eLF7/o9KJoB5GsUzQAcRdGc6V6L5r8KZpoBAAAAA8w0AwAAOJF73YyHnEGnGQAAADBA0QwAAAAYYDwDAADAiVhNfox2fkWnGQAAADBApxkAAMCJsBHQHHSaAQAAAAMUzQAAAIABxjMAAACciDWDjYBmoNMMAAAAGKDTDAAA4ESsVrMzyJ/oNAMAAAAGKJoBAAAAA4xnAAAAOBE2ApqDTjMAAABggE4zAACAE6HTbA46zQAAAIABimYAAADAwD2PZ9y8eVMFCxaUJF29elVbtmyRi4uLWrVqpYceeijHEwQAAMD/cE6zORwumhMSEjRkyBAlJCRo/fr1SkxMVIcOHRQXFyer1aq5c+dq9erVevTRR3MzXwAAAOCBc3g8Y9asWfr+++/19NNPS5I2bNigc+fO6Y033tDy5cvl4uKiWbNm5VqiAAAAyNwIaOZPfuVwp3n79u3q0aOHXn31VUnSV199pSJFiqhv376SpO7du2vJkiW5kyUAAABgIoc7zZcvX1aFChUkSdevX9dPP/2kevXq2a77+fnp5s2bOZ8hAAAAYDKHO80BAQE6ffq0pMwuc3p6uho1amS7vn//fj388MM5niAAAAD+x2rNvyMSZnK4aG7cuLGWLVumxMREbd68WYULF1aTJk104cIFLVy4UJs2bdLLL7+cm7kCAAAApnC4aH7jjTd08+ZNbdiwQQEBAZowYYI8PT117NgxrVq1Sm3btlVYWFhu5goAAACYwmK13t9pfykpKYqPj1exYsVyKqcckXopxuwUADiBetX6mJ0CACfxw7ldZqcgSfqlcgtTP7/8f7409fPNcs8PN/m91NRUffvtt3J1dZWfn5/c3O7rdgAAAECe5HCVm5KSooiICJ05c0aLFy9WSkqKunTpoiNHjkiSypUrp2XLlqlIkSK5liwAAEB+l8FGQFM4fORcZGSk1q1bZzshY+PGjYqOjlZoaKjeeustXbx4Ue+++26uJQoAAACYxeFO8z//+U917NhRERERkqQvv/xSPj4+GjFihNzc3HT69GmtX78+1xIFAAAAzOJw0Xz+/HlVr15dknTz5k3t3btXjRo1ss0xP/zww0pISMidLAEAACCJc5rN4vB4RtGiRXXp0iVJ0r/+9S+lpKTYPdzk6NGjKl68eI4nCAAAAJjN4U5z7dq1tWzZMhUoUECrVq1SwYIF1axZMyUkJOijjz7SunXr1LVr19zMFQAAIN+zZtBpNoPDRXN4eLguXLigqVOnqlChQoqIiJCvr6/27dunqVOnqlatWho0aFBu5goAAACYwuGi2dfXV0uWLNGVK1fk7e0tDw8PSVKlSpW0du1aBQUF5VqSAAAAgJnu+Wkk/v7+dq8LFSpkK5ivXLmS5ToAAAByzv09yxl/1j0VzRs3btTWrVuVlJSkjIwM23p6erpu3LihX375RYcPH87xJAEAAAAzOVw0L1y4UDNnzpS7u7u8vb119epVlShRQteuXdPNmzfl6emp0NDQ3MwVAAAg32MjoDkcPnLu448/VsWKFfXtt99q7dq1slqtWr58uaKiojRu3DglJycz1wwAAIC/JIeL5rNnz+q5556Tt7e3Hn30URUuXFhRUVFydXVVt27dFBISomXLluVmrgAAAIApHC6a3dzc5OXlZXtdpkwZHT161Pa6du3aOnHiRI4mBwAAAHsZVoupP/mVw0VzuXLl9OOPP9peBwYG2m36S0hIUEpKSs5mBwAAAOQBDhfN7du318cff6zhw4crKSlJTZo0UVRUlCIjI7VlyxYtXbpUFStWzM1cAQAA8j2r1WLqT37l8OkZL7zwgs6fP69Vq1bJzc1NzZs317PPPqvIyEhJkre3t4YPH55riQIAAABmsVit93ZEdlpamtzc/ldr7927V/Hx8QoODlaRIkVyPME/K/VSjNkpAHAC9ar1MTsFAE7ih3O7zE5BknQosI2pn1819jNTP98s9/xEwN8XzJJUq1atHEsGAAAAd8cTAc1xx6K5Z8+e93wzi8XCsXMAAAD4y7lj0XzmzJkHmQcAAAAckJ+PfTPTHYvm7du3P8g8AAAAgDzL4SPnJOncuXN6++23FR8fb1tbuHChpk6dqitXruR4cgAAAEBe4HDRfOzYMbVr105LlixRXFycbT0+Pl6rVq3Sc889p9OnT+dKkgAAAMjEOc3mcLhonjFjhry8vLR582a7h5gMHz5cmzdvlru7u95+++1cSRIAAAAwk8NF808//aRevXrpsccey3Lt0UcfVY8ePbR3796czA0AAADIExw+p9lqtSo5Ofmu12/dupUjSQEAACB7nNNsDoc7zUFBQVq7dq0SEhKyXLtx44bWr1+voKCgHE0OAAAAyAsc7jQPGjRIPXr0UOvWrdWmTRuVKVNGFotFp06d0ubNm3Xx4kVNmTIlN3MFAADI9zin2RwOF81BQUFasmSJpk6dqsWLF8v6u/82ULFiRU2ZMkXBwcG5kiQAAABgJovVeu+TMVeuXNHZs2eVkZGhhx9+WMWLF8+N3O6Lm0dJs1MAAAB/IWkpZ81OQZIUVep5Uz+/5pmNpn6+WRzuNP+ev7+//P39czoXAAAAGMjPZyWb6Z6eCAgAAADkR3+q0wwAAABzsBHQHHSaAQAAAAMUzQAAAICBPzWeceHCBZ0/f15ly5ZVgQIF5ObmJhcX6m8AAIDcxgMBzXFPle6+ffvUvn17NWrUSF27dtXhw4f1ww8/qFGjRtqyZUtu5QgAAACYyuGi+eDBg+rTp49u3LihXr162dYLFy4sNzc3DR8+XLt27cqVJAEAAJApw2ox9Se/crhofvfdd1WqVClt2rRJYWFhticCVq1aVZ9++qnKlSun+fPn51qiAAAAgFkcLpp//PFHtW/fXp6enrJY7H/L8Pb2VufOnXX8+PEcTxAAAAAw2z1tBPTw8LjjteTkZGVkZNx3QgAAALgznghoDoc7zUFBQfr888+zvZaUlKT169eratWqOZYYAAAAkFc43Gl+9dVXFRoaqh49eqhp06ayWCw6ePCgjh8/rhUrVujcuXOaOHFibuYKAACQ7/Hf9c1hsd7e0eeAPXv2aPz48Tpz5ozderFixTR27Fg1b948xxP8s9w8SpqdAgAA+AtJSzlrdgqSpH+V6Gjq5z99foOpn2+We5pprlevnrZt26aff/5Zp0+fVkZGhkqWLKkqVarIze1PPScFAAAAyPPuudK1WCyqUqWKqlSpkhv5AAAA4C6sYiOgGRwumnv27OlQ3PLly/90MgAAAEBe5HDR/Mc5ZknKyMjQ1atXlZycrJIlS6pChQo5mhwAAADsZTi8Gw05yeGiefv27dmup6en6+uvv9aYMWPUr1+/HEsMAAAAyCscPqf5TlxdXdW8eXN16tRJb7/9dk7kBAAAAOQp91003/bYY4/pyJEjOXU7AAAAZCNDFlN/8qscKZpTUlL06aefqkiRIjlxOwAAACBPue/TM1JSUhQbG6uEhAQNHjw4xxIDAAAA8or7Oj1DypxpLlu2rFq3bq1u3brlWGIAAADIinOazeFw0bxhwwb5+/vnZi4AAABAnuTwTHP79u01d+7c3MwFAAAABjJM/smvHC6ar1y5oqJFi+ZmLgAAAECe5HDR3KZNG61du/aOs80AAADAX5XDM80uLi6KiYlRixYtVLp0aRUpUkQuLvY1t8Vi0bJly3I8SQAAAGRiI6A5HC6a9+zZIz8/P0lScnKyzp07l2tJAQAAAHmJw0Xz9u3bczMPAAAAOCA/b8Yz0x1nmkeNGqUDBw48yFwAAACAPOmORfMnn3yiU6dOPchcAAAAgDzJ4fEMAAAAmI/xDHM4fOQcAAAAkF/dtdMcFRWl9PT0e7rh888/f18JAQAA4M44cs4cFqvVas3uQsWKFWWxOP43xWq1ymKxKDo6OseSux9uHiXNTgEAAPyFpKWcNTsFSdLmgBdM/fxnL6wx9fPNctdOc+fOnVW9evUHlQsAAACQJ921aK5Zs6batGnzoHIBAACAgQymM0zBRkAAAADAAEfOAQAAOJEMNgKa4o6d5nbt2ql06dIPMhcAAAAgT7pjp3nKlCkPMg8AAAAgz2I8AwAAwIlke1Ywch0bAQEAAAADFM0AAACAAcYzAAAAnEiG2QnkU3SaAQAAAAN0mgEAAJxIhoVzms1ApxkAAAAwQNEMAAAAGGA8AwAAwIlwTrM56DQDAAAABug0AwAAOBGOnDMHnWYAAADAAEUzAAAAYIDxDAAAACeSwTHNpqDTDAAAABigaAYAAHAiGbKY+nPP+WZkaM2aNWrTpo2Cg4PVrFkzTZkyRYmJibaYQ4cOKTQ0VMHBwapfv75mzpyp1NRUu/ucOHFCAwYMUM2aNVW7dm2NHz/e7h6SdOnSJQ0bNky1a9dWjRo1NHToUF28eNEu5saNG5o4caLq1aun4OBgvfjiizpx4oTh92A8AwAAALnmgw8+0KxZs9SvXz/VqVNHsbGxmj17tn755RctWrRIJ0+eVO/evRUcHKxZs2bp119/1TvvvKPExESNGzdOkhQfH69evXqpWLFimjp1qi5fvqzp06fr/Pnzmj9/viQpLS1N/fr1U1JSkiZMmKC0tDTNmDFD/fv310cffSQ3t8yyd8iQITp06JBGjBghLy8vRUZGqmfPntq8ebN8fHzu+D0omgEAAJArrFarPvjgA3Xp0kXDhg2TJNWtW1d+fn4aMmSIoqOjtXLlSvn4+Gju3Lny8PBQw4YN5enpqYiICL300ksKCAjQqlWrlJCQoI0bN8rPz0+SFBAQoLCwMB04cEBBQUHavHmzjhw5oi1btqhcuXKSpEqVKql169baunWrQkJCFBUVpV27dmnhwoVq0KCBJKlmzZpq2rSp1qxZo7CwsDt+F8YzAAAAnIjV5J97cePGDbVt21atW7e2Wy9btqwk6dSpU9qzZ48aN24sDw8P2/WWLVsqPT1du3fvliTt2bNHtWrVshXMklS/fn15eXlp165dtpjy5cvbCmZJtte/j/Hy8lK9evVsMf7+/qpVq5a++eabu34XimYAAADkCm9vb40ZM0Y1atSwW//qq68kSeXKlVNcXJwCAwPtrvv7+8vb21uxsbGSpJiYmCwxrq6uKlWq1F1jJKl06dJ2MWXKlJGrq+sdY+6E8QwAAAAnYvaRcwkJCUpISMiy7uvrK19fX8P3HzhwQAsWLFCzZs1s8d7e3lnivLy8bBv9rl+/7lBM+fLls405efKkJCkxMdHwPndC0QwAAACHLVu2TJGRkVnWBw0apMGDB9/1vfv27dOAAQNUqlQpRUREKCUlRZJksWT9TcBqtcrF5X9DETkRY7XeecDk9/fJDkUzAAAAHNarVy+1a9cuy7pRl/n/27vzuCjrtY/jXwQBYzFI08KNXFPSUI+eXDA198rMJUuRMtdwK0yx1bJccCncSssMydR8UnPXNPUYmeZWejJxRTFERQEBlWXm+cOHOU2At55HvWf08+41rxfzu39z39eMCZcX1+93r169WpGRkapUqZI+//xz+fn5KTMzU5IKrfJmZWXZdrPw9vYudE5mZqYCAgIM5+RXl729vZWYmHjNOUUhaQYAAHAiFpOvf71tGH81d+5cTZgwQQ0aNNCMGTNsybCXl5fKlClja5/Il5KSooyMDFuPcmBgYIE5eXl5SkxMVJs2bWxz4uPjC1z7xIkTqlOnjm3Otm3bZLVa7arSCQkJhfZD/xULAQEAAHDLLF68WOPHj1e7du30+eefF9gLuXHjxtq0aZOtVUOS1q1bJ1dXVzVo0MA2Z/v27UpNTbXN+fHHH5WVlaVGjRpJurqbxqFDh3T06FHbnMOHD+vo0aN2c9LT0/XTTz/Z5pw/f147d+60zSmKi/VazR1OzM09wOwQAADAHSQ3+5TZIUiS5gb0NPX6L5366rrnpqSkqGXLlvL391dUVJTtBiP5KlSooAsXLqhTp06qW7euwsLCdPz4cU2ZMkWdO3fW6NGjJV1NbNu3b6+yZcsqPDxcqampmjhxourUqaPPPvtMkpSdna2nn35a2dnZioiIkNVq1eTJk+Xt7a2lS5farh0aGqr4+HgNHz5c9957r6ZNm6bU1FStWLFCJUuWLPK9kDQDAABcB5Lmq24kaV62bJlGjhxZ5PGoqCh17NhRO3fuVFRUlA4cOCA/Pz8988wzGjx4sIoXL26bGx8fr7Fjx2rPnj3y8vLSE088oREjRtj1IiclJenDDz9UXFyc3N3d1bhxY0VGRur++++3zUlLS9P48eO1YcMGWSwW1atXT5GRkba9o4tC0gwAAHAdSJqvupGk+U7CQkAAAAAnYvY+zXcrFgICAAAABkiaAQAAAAO0ZwAAADgRs/dpvltRaQYAAAAMUGkGAABwIlSazUGlGQAAADBA0gwAAAAYoD0DAADAiVjZp3ZLAh8AACAASURBVNkUVJoBAAAAA1SaAQAAnAgLAc1BpRkAAAAwQNIMAAAAGKA9AwAAwInQnmEOKs0AAACAASrNAAAATsRqdgB3KSrNAAAAgAGSZgAAAMAA7RkAAABOxMIdAU1BpRkAAAAwQKUZAADAibDlnDmoNAMAAAAGSJoBAAAAA7RnAAAAOBHaM8xBpRkAAAAwQKUZAADAiXBHQHNQaQYAAAAMkDQDAAAABmjPAAAAcCLcEdAcVJoBAAAAAyTNAAAAgAHaMwAAAJwI+zSbg0ozAAAAYIBKMwAAgBNhn2ZzUGkGAAAADJA0AwAAAAZozwAAAHAiFho0TEGlGQAAADBApRkAAMCJsOWcOag0AwAAAAZImgEAAAADtGcAAAA4EZYBmoNKMwAAAGCASjMAAIATYSGgOag0AwAAAAZImgEAAAADtGcAAAA4EYuL2RHcnag0AwAAAAaoNAMAADgRC5vOmYJKMwAAAGCApBkAAAAwQHsGAACAE6E5wxxUmgEAAAADVJoBAACcCHcENAeVZgAAAMAASTMAAABggPYMAAAAJ8I+zeag0gwAAAAYIGkGAAAADNCeAQAA4ERozjAHlWYAAADAAJVmAAAAJ8I+zeag0gwAAAAYIGkGAAAADNCeAQAA4ETYp9kcVJoBAAAAA1SaAQAAnAh1ZnNQaQYAAAAMkDQDAAAABmjPAAAAcCLs02wOKs0AAACAASrNAAAATsTKUkBTUGkGAAAADJA0AwAAAAZozwAAAHAiLAQ0B5VmAAAAwACVZgAAACdiYSGgKag0AwAAAAZImgEAAAADtGcAAAA4EZozzEGlGQAAADBA0gwAAAAYIGnGHa9u8CNau3qB0i4c0vlzf2jZ0i9VrVpls8MCYLLGjf6hTRu/VXrqYSUc26mp0R/qvvv87OZUqlRei7/5TGdO79eZ0/s194tolSrlb1LEwFUWWU193K1crFbrHfnu3dwDzA4BDqBatcravm21srIu6ePo2ZKkV4f1l4uLi+rWb6WkpGSTIwRghmYhj2n1qvlKTU3XjJlzlZeXpyGD++j8hVQ1Demo1NQ0+fv7acfPa+TuXlzTps+Rm5ubIl4boOMJiXqsUQfl5OSY/TZwm+VmnzI7BElS/0pdTb3+rOOLTb2+WVgIiDvakMF95OPjreYtn9Xevf+WJG3aFKeft63WsCF9NXLUByZHCMAMH388Rnl5FjVt1lFHjyZIkpZ9t0Z7dm3QG5FDNCJyjF4d1k/lyj2gR+u21B9/HJYk7dixR+vWLlSv0K6a88XXZr4F3MW4I6A5aM/AHe2hwAo6ezbFljBL0s5dv+rcufMKCqphYmQAzFKxYjk9EvSwvpr/rS1hlqSDB49o5arvFRp6tYrXrevT2rJlmy1hlqSNP2zVHwcP67luHW973ADMRdKMO9qhw8fk73+vXQ+in9+9uvdeXyWdPmNiZADMEvBgWUnS/v0HChw7cuS4Spe+T+XKPajKlStp957fCszZs2ef6tWrfcvjBOBYSJpxR5s0eaYSE5M0P3amHnnkYQUF1dD82BnKzs7R9BlzzA4PgAkysy5Jknx8vAsc8/e/uhCwUsVykqRTp04XmHM66YxKlvSVr6/PLYwSKJrV5P/uViTNuKOdPPmnxk+YppCQf2rPrg3au3ujWrRoop69wu1aNgDcPX7/PV5paenq1Km93biHh4dat3rcbizr/xLsv7p0+bIkycvrnlsWIwDHY9pCwOTkG9u1oEyZMrcoEtzJ3hv9ut58Y5i2bPlJn82ZL1dXVw3o10sLv/5U3Z7rp5Wrvjc7RAC3WU5Ojj6Onq133xmu2HnTNSFqulxdXfX+6BHy8iohSbJYri61utYGU/lzgNuN//PMYVrS3KxZM7m4uFz3/AMHCvaeAddSsqSvIl4boF927lWrNs/ZfsAtWvSdft62Wp9+EqWHqjRUdna2yZECuN0++PBj3VuypAYN6q3nu3eSJK1YuV4TJ83UuLFvKiMzS5JUooRngdeW8Lw6dvFixu0LGIDpTEuax44da0ua09LSNGnSJD322GNq166dSpcurdTUVP3www/avHmzIiMjzQoTTqxqlUB5enpq0aLv7CpCubm5WrBgiSaMf1s1alTRb7/9bmKUAMxgtVoV8fpoTZg4XVWrBOpk4p86ceKUxrw/Urm5uUpOPitJeuCBgr/lfODBMrpwIbXQ1g0Ady7TkuZnn33W9nV4eLg6deqkMWPG2M156qmn9MEHH2jNmjV67rnnbneIcHJX/q+C7OpasHXf1dVVklSs2PX/tgPAneO55zrqdNIZbfnXNp05c8423rRJQ+3e/ZuSk8/q6NEEBT8aVOC1jz4apF27Cu6qAdwud/NiPDM5xELAuLg4tW3bttBjzZs31549e25zRLgT/PvfB3XqVJJ69eomDw8P27iHh4d69uyis2dTtH//QRMjBGCWYUP6Kjr6A9s/oCWpfbuWatKkoWZ+GiNJWrp0tVq2bKrq1Svb5rRs0VQ1qlfRom++u+0xAzCXQ9wR0M/PT7/99psaN25c4NiOHTtYBIj/isVi0ZBhb+mbhbO17aeVmjt3oVxdXfXii8+pRvUqCntpiHJzc80OE4AJoibN1OJFn2n5shgtXbZGlSqV17ChfbVu3SZ9/fUSSdLEyTPVs2cXrV+7SB99PFuenh4aHjFQO3f9qvn/NwcwAwsBzeFivdbS4Ntk5syZmjlzpl5++WW1bNlSfn5+SklJ0dq1axUbG6s33nhDPXr0uKFzurkH3KJo4WyaP95Yb705TPXrPyrp6o0Jxo2fqnXrN5sbGABTdev2tEa8PkhVqwQqOfmsFixcqvETpunSpcu2OdWqVdbkie+qadN/Kivrktas/UEjI8fo3LnzJkYOs+RmnzI7BElSWKXOpl4/5vi3pl7fLA6RNFutVkVFRSk2NlZ5eXm2cQ8PD73yyivq16/fDZ+TpBkAANxMJM1XkTQ7gIsXL2rPnj1KT0+Xn5+fgoODdc89/93m8STNAADgZnKUpDm04rPGk26h2IS7sz3JIXqa8/n4+CgkJMTsMAAAAAA7DpE0t27d2vBGJ+vWrbtN0QAAADguh2kRuMs4RNJct27dAklzZmam9u3bpytXrigsLMykyAAAAAAHSZrHjx9f6HhOTo5eeeUVXbrEXZcAAABgHoe4uUlRihcvrl69eul//ud/zA4FAADAIVhkNfVxt3LopFmS0tLSlJmZaXYYAAAAuIs5RHvGihUrCozl5eXp9OnTiomJUf369U2ICgAAwPFY7+Jqr5kcIml+/fXXizwWHByst99++zZGAwAAANhziKR548aNBcZcXFzk7e0tX19fEyICAAAA/sMhkuaAgGvfvS8zM1NeXl63KRoAAADHZTE7gLuUQyTN2dnZio2N1S+//KKcnBzl39nbYrHo0qVLOnjwoPbu3WtylAAAALhbOUTSPGnSJM2bN0/VqlXT+fPn5eHhIX9/f8XHxysnJ0eDBg0yO0QAAADcxRxiy7l169bppZde0vLly9WzZ08FBQVp8eLFWr9+vQICAmSx8IsIAAAAiX2azeIQSXNKSopCQkIkSdWqVdO+ffskSWXKlFG/fv20evVqM8MDAADAXc4h2jN8fHyUk5MjSapYsaKSkpKUkZEhb29vVapUSUlJSSZHCAAA4BjYp9kcDlFprlevnr766itdvnxZFStWVIkSJbRhwwZJ0q+//ipvb2+TIwQAAMDdzCGS5vDwcO3atUt9+/aVm5ubXnjhBb3zzjvq2rWrPvroI7Vp08bsEAEAAHAXc4j2jIcfflirV69WfHy8JCkiIkLe3t7avXu3Bg4cqH79+pkcIQAAgGNgewRzOETSPGHCBLVv315NmzaVdPVugAMGDDA5KgAAAOAqh2jP+Oabb5Senm52GAAAAA7ParWa+rhbOUTSXKtWLf30009mhwEAAIBb6MCBA6pVq5ZOnz5tN/7jjz+qc+fOqlOnjlq0aKEvvviiwGv37dun0NBQBQcHq0mTJpoyZYpt97V8x48f14ABA1S/fn01bNhQ7777rjIyMuzmnDt3ThEREWrYsKHq1aun1157TWfPnjWM3SHaM2rVqqWYmBitX79eVapUUalSpeyOu7i46P333zcpOgAAAPx/HT16VP3791dubq7d+O7duzVgwAC1a9dOQ4cO1a5duxQVFSWr1aqXX35ZkpSQkKAXX3xRwcHB+vjjj3XkyBF99NFHysjI0DvvvCNJSktLU1hYmEqXLq0JEyYoJSVFEydO1OnTpzVr1ixJUm5url5++WVlZWVp9OjRys3N1eTJk9WnTx99++23cnMrOjV2iKR53bp1uv/++5WXl6eDBw/q4MGDdsddXFxMigwAAMCxONtd+XJzc7Vo0SJNnjxZxYsXL3B86tSpqlmzpiZOnChJCgkJUW5urj799FOFhobK3d1ds2fPlo+Pj2bOnCl3d3c1a9ZMnp6e+uCDD9S/f3+VKVNG8+fPV3p6upYtWyY/Pz9J/7lR3q+//qo6depo1apV+uOPP7R69WpVrlxZ0tUNKZ588kmtX79e7du3L/J9mNaesW3bNmVmZkqSfvjhh2s+Nm7caFaYAAAA+H/YtWuXJk2apN69e2v48OF2x65cuaKdO3eqdevWduNt2rRRenq6du/eLUmKi4tT8+bN5e7ubpvTtm1b5eXl6ccff7TN+cc//mFLmCWpSZMm8vLy0pYtW2xzqlSpYkuYJdme588pimlJc+/evXXkyBG7sUWLFunChQsmRQQAAOD4LCY/0tPTlZiYWOBR1KYOlStX1oYNGzRo0CC5urraHTt58qRycnIUGBhoN16xYkVJ0rFjx3Tp0iUlJSUVmOPv7y9vb28dO3ZM0tX2j7/PcXV1Vbly5a45R5IqVKhgm1MU09oz/r76Mi8vT6NHj1ZQUJDdvxAAAADgOGJiYjR9+vQC44MGDdLgwYMLjP99rdpfXbx4UZIK3P3Zy8tLkpSRkVHknPx5+Qv9Ll68eF1zqlSpUuichISEIuOUHKSnOd/dvI0JAACAMwgLC1OnTp0KjPv6+t7wufJzv6LWrxUrVuyac6xWq4oV+0/jxM2aUxiHSpoBAABwbVaTFwL6+vr+VwlyYXx8fCSpwLZw+c99fHxs1eO/z5GkrKws2zm8vb0LnZOZmamAgADDOYVVqf/KIfZpBgAAwN2nQoUKcnV11YkTJ+zG858HBgbKy8tLZcqUKdA+kZKSooyMDFuPcmBgYIE5eXl5SkxMvOac/OsV1uv8Vw6XNLO9HAAAQNEsspr6uJk8PDxUv359rV+/3q5Nd926dfLx8VFQUJAkqXHjxtq0aZOys7Pt5ri6uqpBgwa2Odu3b1dqaqptzo8//qisrCw1atRI0tXdNA4dOqSjR4/a5hw+fFhHjx61zSmKqe0ZQ4cOtds6RJLCw8MLjElXPxgAAADcWQYOHKiXXnpJr776qjp16qQ9e/Zozpw5ioiIUIkSJSRJffr00apVq9SvXz+FhYXp+PHjmjJlirp166YHH3xQkvTCCy/oq6++0osvvqjw8HClpqZq4sSJCgkJUd26dSVJ7du316effqo+ffooIiJCVqtVkydPVtWqVdWuXbtrxuliNWn13ahRo25o/rhx425ovpt7wA3NBwAAuJbc7FNmhyBJal+h6Btw3A6rT6z+r1+7ZMkSjRo1Slu2bFHZsmVt499//72mTp2qY8eOqUyZMurRo4d69+5t99qdO3cqKipKBw4ckJ+fn5555hkNHjzY7oYp8fHxGjt2rPbs2SMvLy898cQTGjFihF2/clJSkj788EPFxcXJ3d1djRs3VmRkpO6///5rxm5a0nyrkTQDAICbyVGS5nblr10RvdXWnFxj6vXN4nA9zQAAAICjYcs5AAAAJ2IxO4C7FJVmAAAAwABJMwAAAGCA9gwAAAAnYvYdAe9WVJoBAAAAAyTNAAAAgAHaMwAAAJzIzb6VNa4PlWYAAADAAJVmAAAAJ3KH3szZ4VFpBgAAAAyQNAMAAAAGaM8AAABwIiwENAeVZgAAAMAAlWYAAAAnwh0BzUGlGQAAADBA0gwAAAAYoD0DAADAiVjYp9kUVJoBAAAAA1SaAQAAnAh1ZnNQaQYAAAAMkDQDAAAABmjPAAAAcCLcEdAcVJoBAAAAA1SaAQAAnAiVZnNQaQYAAAAMkDQDAAAABmjPAAAAcCJW7ghoCirNAAAAgAGSZgAAAMAA7RkAAABOhN0zzEGlGQAAADBApRkAAMCJWKk0m4JKMwAAAGCApBkAAAAwQHsGAACAE2GfZnNQaQYAAAAMUGkGAABwImw5Zw4qzQAAAIABkmYAAADAAO0ZAAAAToSFgOag0gwAAAAYoNIMAADgRFgIaA4qzQAAAIABkmYAAADAAO0ZAAAATsRKe4YpqDQDAAAABqg0AwAAOBELW86ZgkozAAAAYICkGQAAADBAewYAAIATYSGgOag0AwAAAAaoNAMAADgRFgKag0ozAAAAYICkGQAAADBAewYAAIATYSGgOag0AwAAAAZImgEAAAADtGcAAAA4EXbPMAeVZgAAAMAAlWYAAAAnwkJAc1BpBgAAAAyQNAMAAAAGaM8AAABwIiwENAeVZgAAAMAAlWYAAAAnwkJAc1BpBgAAAAyQNAMAAAAGaM8AAABwIlarxewQ7kpUmgEAAAADVJoBAACciIWFgKag0gwAAAAYIGkGAAAADNCeAQAA4ESs3BHQFFSaAQAAAANUmgEAAJwICwHNQaUZAAAAMEDSDAAAABigPQMAAMCJsBDQHFSaAQAAAANUmgEAAJyIhUqzKag0AwAAAAZImgEAAAADtGcAAAA4ESv7NJuCSjMAAABggKQZAAAAMEB7BgAAgBNhn2ZzUGkGAAAADFBpBgAAcCIWFgKagkozAAAAYICkGQAAADBAewYAAIATYSGgOag0AwAAAAaoNAMAADgRC5VmU1BpBgAAAAyQNAMAAAAGaM8AAABwIiwENAeVZgAAAMAAlWYAAAAnwh0BzUGlGQAAADBA0gwAAAAYoD0DAADAibAQ0BxUmgEAAAADVJoBAACcCHcENAeVZgAAAMAASTMAAABggPYMAAAAJ2Jln2ZTUGkGAAAADJA0AwAAAAZozwAAAHAi7J5hDirNAAAAgAEqzQAAAE6EOwKag0ozAAAAYICkGQAAADBAewYAAIATYZ9mc1BpBgAAAAxQaQYAAHAiLAQ0B5VmAAAAwABJMwAAAG6plStXqkOHDqpdu7batWunZcuWmR3SDaM9AwAAwIk4W3vGmjVrNHz4cPXq1UtNmzbVhg0bNHLkSHl6eqpt27Zmh3fdXKzO9slfJzf3ALNDAAAAd5Dc7FNmhyBJKm5yjpNzg59Dq1atFBQUpI8++sg2NmzYMB08eFBr1qy52eHdMrRnAAAAOBGryY8bcfLkSZ04cUKtW7e2G2/Tpo2OHj2qkydP3uAZzUN7BgAAAK5benq60tPTC4z7+vrK19fXbuzo0aOSpMDAQLvxihUrSpKOHTum8uXL36JIb647Nml2lF+hAAAA3Exm5zjTpk3T9OnTC4wPGjRIgwcPthu7ePGiJMnb29tu3MvLS5KUkZFxi6K8+e7YpBkAAAA3X1hYmDp16lRg/O9VZuk/ixZdXFwKHS9WzHk6hUmaAQAAcN0Ka8Moio+Pj6SCFeXMzEy7487AedJ7AAAAOJX8XuYTJ07YjSckJNgddwYkzQAAALglKlasqHLlymnt2rV24+vXr1elSpX04IMPmhTZjaM9AwAAALdMeHi4Ro0apZIlS+rxxx/XDz/8oDVr1tjt2+wM7tibmwAAAMAxLFy4UF988YWSkpJUvnx59evXT88884zZYd0QkmYAAADAAD3NAAAAgAGSZgAAAMAASTMcQmhoqGrVqqUDBw4UerxmzZqaNm3abY7q+i1ZskTVq1fX6dOnzQ4FuCtFRkaqevXqRT5++uknU+LiewNw52D3DDiM3NxcvfHGG1q8eLHc3PhfE8CNKVu2rKKjows9VqVKldscDYA7DZkJHIaPj49+//13ffbZZxo4cKDZ4QBwMu7u7nr00UfNDgPAHYr2DDiMoKAgdejQQTNnztSRI0eKnHf58mVFR0erTZs2euSRR9S+fXstWrTIbk6LFi00fvx4hYaGqm7duho3bpy2b9+u6tWra9u2bXrhhRdUu3ZttW7dWhs2bNDRo0cVFhamOnXqqFWrVlq1apXd+datW6fnn39ewcHBCgoKUrt27fT111/fks8BwK1R2PcFSTpw4IDCw8P1z3/+U7Vq1VJISIg+/PBDXblyRZKUmJio6tWr67vvvrM7X2RkpFq1amV7brFYNHPmTD3++OOqU6eOXnnlFaWlpd2+NwjgliJphkN566235OXlpTfeeEMWi6XAcavVqr59+yomJkbPP/+8PvnkEzVq1EjvvvuuZsyYYTc3NjZWQUFBio6OVocOHWzjw4cPV/v27fXJJ5/I19dXI0aM0IABA/T4448rOjpapUuXVmRkpJKTkyVJGzdu1JAhQ1S7dm3NnDlT06ZNU7ly5fTee+/pt99+u7UfCIAbkpubW+Dx151V//59ITk5WT169NCVK1c0YcIEffbZZ2rfvr3mzZunefPm3dC1J06cqBkzZqhLly6aPn26/Pz8NHny5Jv9FgGYhPYMOBR/f3+9/fbbeu211xQTE6OXXnrJ7viWLVu0Y8cORUdHq23btpKkJk2aKDc3V59++qleeOEF+fn5Sbra3zhixAi5uLhIkrZv3y5J6t69u3r27ClJysjI0JAhQxQWFma7VqlSpdS5c2f9/vvvKlOmjI4cOaJnn31Wo0aNssURHByshg0baseOHapdu/at/VAAXJcTJ06oVq1aBcZHjx6t559/XlLB7wv/+te/VLNmTUVHR8vLy0uS1KhRI8XFxemXX35R3759r+va6enpio2NVe/evTVo0CBJUtOmTZWcnKytW7fejLcHwGQkzXA4HTp00MqVKxUdHa2WLVuqQoUKtmO//PKLihcvrtatW9u95qmnntKCBQv066+/6vHHH5ckVa1a1faD8a/+muTed999kmTXB3nvvfdKuvpDUJL69esnScrMzNSxY8d04sQJ7du3T5KUk5Pz/327AG6SsmXLavr06QXGAwICbF///ftCSEiIQkJClJOTo8OHDyshIUHx8fE6f/68SpUqdd3X3rt3r3JyctSyZUu78Xbt2pE0A3cIkmY4pNGjR+vJJ5/Um2++afcr0rS0NN13330qVsy+syj/h9vFixdtY/kJ8d/lV5P+ytPTs8hYzp8/r3fffVcbNmyQi4uLKlasqHr16kmSuKEm4Djc3d31yCOPXHPO378vWCwWTZkyRfPnz1dWVpYeeOAB1a5dWx4eHjf09zu/d9nf399uvHTp0td9DgCOjaQZDqlMmTIaOXKk3nzzTS1YsMA27uvrq5SUFFksFrvE+ezZs5Jka824mYYPH65jx47pyy+/VHBwsNzd3XXp0iUtXrz4pl8LwO01e/Zsffnll3r//ffVqlUr+fj4SJK6dOlim5Nfmf77OousrCzb1/nfe86dO2f327HU1NRbFjuA24uFgHBYXbp0UePGjTVp0iTbD6sGDRooJydH69evt5u7cuVKFS9e/Jb0F+/atUtt27ZVw4YN5e7uLulqH6RU8IcoAOeya9cuVa9eXc8++6wtYU5OTlZ8fLzt77e3t7ckKSkpyfa6nJwcu4XAwcHB8vT01Nq1a+3Ov2nTplv9FgDcJlSa4dDGjBmjJ5980vZr0pCQEP3jH//Qm2++qdOnT6tq1arasmWLFi5cqIEDB8rX1/emx1C7dm0tX75cDz/8sMqUKaPdu3dr9uzZcnFx0aVLl2769QDcPvm74nz22WeqU6eOEhISNGvWLGVnZ9v+fpcsWVLBwcGKiYlR+fLlVbJkSc2bN0+XL19W8eLFJV1t+3rllVf08ccfy9PTUw0aNNDmzZtJmoE7CEkzHFpAQIAiIiI0ZswYSVKxYsU0a9Ysffzxx/r888+VlpamSpUqafTo0erevfstiWH8+PEaM2aM3n//fUlSpUqV9N5772n58uXatWvXLbkmgNujf//+unDhgmJiYnTx4kU98MAD6tixo1xcXDR79mxlZGTI29vb9n3grbfekre3t7p06aJ69eppyZIldue65557FBMTo7lz5yo4OFgjR47U6NGjzXuDAG4aFysrmQAAAIBroqcZAAAAMEDSDAAAABggaQYAAAAMkDQDAAAABkiaAQAAAAMkzQAAAIABkmYAdiIjI1W9enW7x8MPP6y6deuqa9euWrp06W2Jo0WLFgoNDbU9Dw0NVYsWLW74PBkZGTp//vxNiyv/83GU8wAAbg9ubgKgUKNGjZKfn58kyWq1KiMjQ8uXL1dkZKQuXLig3r1739Z4BgwYcMN3YNy/f78GDhyoSZMmqWHDhrcoMgDA3YCkGUChnnjiCZUrV85urEuXLmrfvr1mzJihnj17yt3d/bbF07hx4xt+TXx8vM6cOXMLogEA3G1ozwBw3Tw9PdWiRQtlZGTo0KFDZocDAMBtQ6UZwA1xcXGRJOXl5Um62nvcqFEjWSwWrVixQn5+flq2bJn8/f21Z88eTZ06VXv37pUkBQcHa9iwYapdu7bdOVevXq1Zs2bp2LFjqlChgt56660C1w0NDdWpU6f0ww8/2MaOHDmiqVOn6ueff1Zubq5q1qypoUOHqn79+po2bZqmT58uSerVq5cCAgJsrz19+rSmTJmif/3rX8rMzFTlypXVu3dvPf3003bX3L9/v6ZMmaI9e/bI29tbffv2ve7PKTk5WdHR0bZrPPTQQxo4cKCeeOKJIl+zbds2zZkzR/v27VNGRob8/f3VvHlzDR8+XL6+vpKutsrMmDFDK1as0J9//ikfHx81btxYr732mh544AHbuRYsWKCvv/5aJ06ckKenp+rX/xfwrwAACd1JREFUr69hw4apatWq1/0eAAD/QdIM4LpZLBbt2LFD7u7uqly5sm181apVCgwM1Jtvvqlz587J399fcXFx6t+/v2rUqKGhQ4cqOztbS5YsUY8ePTR37lzVr19fkrRkyRKNGjVKwcHBev3115WQkKABAwbIYrEoICCgyFiOHz+ubt26yc3NTT179pS/v78WLlyol156SfPnz1erVq109uxZLVq0SAMGDNAjjzwi6Woy27VrV1mtVoWGhqpkyZLauHGjXn/9dZ05c0Z9+vSRJB06dEihoaHy9fXVK6+8opycHM2YMcP2j4VrSU1NVbdu3ZSamqoePXqofPnyWrlypQYNGqTp06cXmjj/+OOP6tu3r+rWrashQ4bIxcVFcXFxWrRokXJycjRu3DhJ0qeffqoZM2aoR48eql69uhITEzVv3jzt379fK1eulKurq5YvX67Ro0frmWeeUWhoqM6fP6+YmBiFhobq+++/l4+Pz/X/oQMArrICwF+MHDnSWq1aNeu///1va0pKijUlJcV65swZ6549e6xDhw61VqtWzTp27Fjb/ObNm1tr1KhhTUhIsI3l5eVZW7Zsae3evbs1NzfXNp6ZmWlt1aqVtWPHjlar1WrNzc21PvbYY9bOnTtbs7OzbfO+/fZba7Vq1aw9e/a0jfXs2dPavHlz2/OhQ4daa9eubT1+/Lht7Pz589Z69epZhwwZYneen3/+2e79NWjQwJqcnGz3vl977TVrUFCQ9dy5c1ar1WodPHiw9dFHH7X++eeftjmHDx+2BgUFWatVq3bNzzAqKsparVo1686dO21jly9ftj7xxBPWzp07233O+V5++WVr8+bNrVeuXLE7V7du3azBwcG25+3atbP269fPbs6CBQusTz/9tO3PoE+fPtYOHTrYzdm8ebO1ffv2djEBAK4flWYAherUqVOBMXd3d4WGhioiIsJuvEKFCqpQoYLt+e+//66TJ0/q+eefV1pamt3c5s2b68svv9Tp06d15swZpaSkaNCgQSpevLhtTseOHTV+/PgiY7NYLNqyZYuaNWumihUr2sb9/Pz09ddf23b9KOx1GzZsUMOGDeXm5ma3FV3r1q21cuVKxcXF6cknn9TWrVvVrFkzu5aHypUrq0mTJnYtIoXZvHmzatWqpXr16tnGPDw8NHv2bHl4eBT6mlmzZik9Pd1uceWFCxfk7e2trKws21jZsmW1fft2xcTEqEOHDipVqpS6d++u7t27282Ji4vT9OnT9cwzz6hcuXJq1qyZmjVrds24AQBFI2kGUKiJEyeqVKlSkqRixYrJ19dXlStXLjTpu+++++yenzhxQpIUFRWlqKioQs+flJSk06dPS5Jdwi1Jrq6udsnw36WmpiorK6vQOdWqVSvydRcuXNDFixe1YcMGbdiwoci48s//97gk6aGHHjJMmk+dOlXontKBgYFFvsbV1VUnT55UdHS0Dh8+rBMnTig5ObnAvBEjRmjgwIEaO3asxo0bp1q1aqlFixbq1q2bSpcuLUkKDw/X3r17NW3aNE2bNk1VqlRRixYt1LVr10LfEwDAGEkzgELVrVu3wJZzRXF1dbV7brFYJElDhw7Vo48+WuhrHnroIVtSeOXKlQLH889RmPy+4mLFbmwDoPzXtWnTxq4y+1fly5e3fX2jcf31OvkLJq/XwoUL9e677yowMFD169dX69atVadOHcXGxmrFihW2eTVq1NC6deu0detWbdq0SVu3btXUqVP15ZdfauHChapcubLKli2r7777Ttu3b9fGjRu1detWzZ49W3PnztUXX3yhBg0a3FBsAACSZgC3QP4CvnvuuUeNGjWyO/bbb78pLS1Nnp6etgT1+PHjdnOsVqtOnTpV5E4Pfn5+8vT0VEJCQoFjc+bM0blz5zRy5MgCx/z9/VWiRAnl5uYWiOvPP//U77//rhIlSsjPz0/e3t4F4pKkxMTEIt93vgcffNBWbf+rpUuXateuXXrnnXfsxq9cuaLx48erYcOG+uKLL+Tm9p9vzdHR0bav8/Ly9Mcff8jb21stW7ZUy5YtJV3dfeTVV1/V4sWLFRkZqYMHD0qSHnvsMT322GOSpF27diksLEyxsbEkzQDwX2CfZgA3XVBQkEqXLq3Y2FhlZmbaxjMyMjRs2DCNGjVKrq6uqlmzpgICArRgwQK7u/2tWrVKFy5cKPL8bm5uaty4sbZs2aKkpCTbeFpamubMmWNLWPMr0fnVYTc3N4WEhGjLli36448/7M45fvx4hYeH68KFC3JxcVGrVq20detWxcfH2+YkJiZq8+bNhu8/JCRE+/bt0/79+21jOTk5mjNnjvbv31/gpjCXL1/WpUuXVKlSJbuE+cCBA9qxY4ckKTc3V3l5eerVq5fGjh1r9/o6derYvd+hQ4dqxIgRdjt91KxZU8WLF7/h6jwA4CoqzQBuuuLFi+vtt9/WsGHD9Oyzz6pLly7y8PDQ4sWL9eeff2rSpEm25PDtt99WeHi4nnvuOXXu3FnJycmaP3++7r333mteIyIiQl27dlXXrl3Vo0cPeXt765tvvlFWVpaGDRsm6WplWbq6Z/G5c+f01FNPafjw4dq+fbt69OihHj166MEHH9TmzZu1adMmPffcc7bq9tChQ7V582aFhobqxRdflKurq2JjY+Xl5aXs7Oxrxta/f3+tXbtWYWFh6tmzp+6//36tWrVKR44c0Zw5cwrML1mypOrUqaMlS5bI29tbgYGBOnTokBYvXmxLcjMzM1WyZEmFhobqk08+UXh4uJo2barLly9r0aJFKlGihDp37ixJevnll/XWW2/pxRdfVNu2bWW1WvXdd9/pypUreuGFF27gTxIAkM/FarVazQ4CgOOIjIzU0qVLtXHjxuvqaW7RooUCAgIUGxtb4Ni2bdv0ySefaN++fSpWrJiqVq2q/v37q3nz5nbztm7dqmnTpungwYMqU6aMhgwZovnz58vNzc123sJubnLo0CFNmTJFO3bsULFixVS7dm1FRESoZs2akq5Wd19//XVt2rRJHh4e2rp1qzw8PJSQkKCpU6cqLi5OWVlZKl++vLp27arQ0FC7/uxjx44pKirKtjd1165dJV3d6SK/BaIoSUlJmjx5srZu3ars7GzVqFFDQ4YMsbVL5H/O+edJSkrSuHHjtH37dmVnZysgIEAdOnRQ5cqVNXjwYE2dOlVt2rSRxWLRvHnz9O233yoxMVGurq62vZ2DgoJs11+2bJnmzZunhIQEWSwWBQUFacCAAf/V7cgBACTNAAAAgCGa2wAAAAADJM0AAACAAZJmAAAAwABJMwAAAGCApBkAAAAwQNIMAAAAGCBpBgAAAAyQNAMAAAAGSJoBAAAAAyTNAAAAgIH/BdpOCjbswphtAAAAAElFTkSuQmCC\n",
+      "text/plain": [
+       "<Figure size 864x864 with 2 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "# Confusion matrix\n",
     "\n",
@@ -4990,6 +5163,14 @@
     "plt.show()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "**Autoencoder 모델은 Fraud 거래를 매우 편향된 데이터 속에서도 탐지할 수 있습니다.**  \n",
+    "하지만 사기거래를 대부분 잡더라도, 보통(Normal) 거래도 사기로 12% 정도나 판단합니다! 더 큰 임계값(threshold)는 miss 분류를 줄일 수는 있겠으나, 사기 클래스의 탐지 정확도를 떨어트리게 됩니다."
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {
@@ -4998,9 +5179,6 @@
     "collapsed": true
    },
    "source": [
-    "** Autoencoder model is able to detect Fraud transaction in this highly skewed data!**  \n",
-    "Although the model catches most of Fraud transaction, it also classifies lot of normal transaction (12%!) as Fraud. Larger threshold will reduce misclassification of the Normal class but it will also reduce detection of the Fraud class.\n",
-    "\n",
     "## 3. Visualizing the Data with t-SNE\n",
     "\n",
     "**t-Distributed Stochastic Neighbor Embedding (t-SNE)** is a technique ([wiki page](https://en.wikipedia.org/wiki/T-distributed_stochastic_neighbor_embedding)) for dimensionality reduction that is particularly well suited for the visualization of high-dimensional datasets. It was developed by Geoffrey Hinton and Laurens van der Maaten.  \n",