-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy patht1_t2_ann.R
448 lines (299 loc) · 21.3 KB
/
t1_t2_ann.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
library(data.table)
returnUnixDateTime<-function(date) {
returnVal<-as.numeric(as.POSIXct(date, format="%Y-%m-%d", tz="GMT"))
return(returnVal)
}
findTimeToNearestHbA1c <- function(dmList_LinkId, diagnosisDate_unix) {
hb_id_sub <- cleanHbA1cDataDT[LinkId == dmList_LinkId]
diffFromDiagnosisDate <- hb_id_sub$dateplustime1 - diagnosisDate_unix
findClosest <- sqrt(diffFromDiagnosisDate ^ 2)
flagClosest <- ifelse(findClosest == min(findClosest), 1, 0)
outputList <- list(findClosest[flagClosest == 1], hb_id_sub$hba1cNumeric[flagClosest == 1])
return(outputList)
}
# set index date
index <- "2017-01-01"
paramFromDiagnosisWindowMonths = 12
paramFromDiagnosisWindowSeconds = paramFromDiagnosisWindowMonths * (60*60*24*(365.25/12))
numberOfYearsData <- 1 # minimum length of time within the dataset to allow correct diagnosis to have been reached
# diagnosisDataset<-read.csv("../GlCoSy/SDsource/diagnosisDateDeathDate.txt")
diagnosisDataset<-read.csv("~/R/GlCoSy/SDsource/demogALL.txt", quote = "",
row.names = NULL,
stringsAsFactors = FALSE)
diagnosisDatasetDT = data.table(diagnosisDataset)
# cut down dataset
cut_diagDT <- data.table(diagnosisDatasetDT$LinkId, diagnosisDatasetDT$DateOfDiagnosisDiabetes_Date, diagnosisDatasetDT$Ethnicity_Mapped, diagnosisDatasetDT$CurrentGender_Mapped, diagnosisDatasetDT$DiabetesMellitusType_Mapped, diagnosisDatasetDT$BirthDate)
colnames(cut_diagDT) <- c("LinkId", "diagnosisDate", "Ethnicity", "Sex", "diabetesType", "DOB")
cut_diagDT$LinkId <- as.numeric(cut_diagDT$LinkId)
cut_diagDT$LinkId <- as.numeric(cut_diagDT$LinkId)
cut_diagDT$DOB_unix <- returnUnixDateTime(cut_diagDT$DOB)
# remove those without a diagnosis date
cut_diagDT$diagnosisDate_unix <- returnUnixDateTime(cut_diagDT$diagnosisDate)
cut_diagDT$diagnosisDate_unix[is.na(cut_diagDT$diagnosisDate_unix)] <- 0
cut_diagDT <- cut_diagDT[diagnosisDate_unix > 0]
# flag first diagnosis and remove all others from dataset
cut_diagDT[, c("firstDiagnosis") := ifelse(diagnosisDate_unix == min(diagnosisDate_unix), 1, 0) , by=.(LinkId)]
cut_diagDT <- cut_diagDT[firstDiagnosis == 1]
# cut to T1 or T2
cut_diagDT <- cut_diagDT[diabetesType == "Type 1 Diabetes Mellitus" | diabetesType == "Type 2 Diabetes Mellitus"]
# cut to ensure at least 1y data for each ID
cut_diagDT <- cut_diagDT[diagnosisDate_unix < (returnUnixDateTime(index) - ((60*60*24*365.25) * numberOfYearsData))]
# generate node and link files
cleanHbA1cData <- read.csv("~/R/GlCoSy/SD_workingSource/hba1cDTclean.csv", sep=",", header = TRUE, row.names = NULL)
cleanHbA1cData$timeSeriesDataPoint <- cleanHbA1cData$hba1cNumeric
cleanHbA1cDataDT <- data.table(cleanHbA1cData)
hb_DT_forMerge <- data.table(cleanHbA1cDataDT$LinkId, cleanHbA1cDataDT$dateplustime1, cleanHbA1cDataDT$hba1cNumeric)
colnames(hb_DT_forMerge) <- c("LinkId", "hb_dateplustime1", "hba1cNumeric")
cleanSBPData <- read.csv("~/R/GlCoSy/SD_workingSource/SBPsetDTclean.csv", sep=",", header = TRUE, row.names = NULL)
cleanSBPData$timeSeriesDataPoint <- cleanSBPData$sbpNumeric
cleanSBPDataDT <- data.table(cleanSBPData)
sbp_DT_forMerge <- data.table(cleanSBPDataDT$LinkId, cleanSBPDataDT$dateplustime1, cleanSBPDataDT$sbpNumeric)
colnames(sbp_DT_forMerge) <- c("LinkId", "sbp_dateplustime1", "sbpNumeric")
cleanDBPData <- read.csv("~/R/GlCoSy/SD_workingSource/DBPsetDTclean.csv", sep=",", header = TRUE, row.names = NULL)
cleanDBPData$timeSeriesDataPoint <- cleanDBPData$dbpNumeric
cleanDBPDataDT <- data.table(cleanDBPData)
dbp_DT_forMerge <- data.table(cleanDBPDataDT$LinkId, cleanDBPDataDT$dateplustime1, cleanDBPDataDT$dbpNumeric)
colnames(dbp_DT_forMerge) <- c("LinkId", "dbp_dateplustime1", "dbpNumeric")
cleanBMIData <- read.csv("~/R/GlCoSy/SD_workingSource/BMISetDTclean.csv", sep=",", header = TRUE, row.names = NULL)
cleanBMIDataDT <- data.table(cleanBMIData)
bmi_DT_forMerge <- data.table(cleanBMIDataDT$LinkId, cleanBMIDataDT$dateplustime1, cleanBMIDataDT$bmiNumeric)
colnames(bmi_DT_forMerge) <- c("LinkId", "bmi_dateplustime1", "bmiNumeric")
# cleanRenalData <- read.csv("~/R/GlCoSy/SD_workingSource/renalSetDTclean.csv", sep=",", header = TRUE, row.names = NULL)
# cleanRenalDataDT <- data.table(cleanRenalData)
# renal_DT_forMerge <- data.table(cleanRenalDataDT$LinkId, cleanRenalDataDT$dateplustime1, cleanRenalDataDT$egfrNumeric)
# colnames(renal_DT_forMerge) <- c("LinkId", "egfr_dateplustime1", "egfrNumeric")
# find closest value to diagnosis date for each parameter - sequential merge
# hba1c
merge_hb <- merge(hb_DT_forMerge, cut_diagDT, by = "LinkId")
merge_hb[, c("hb_diffFromDiag") := sqrt((hb_dateplustime1 - diagnosisDate_unix) ^ 2) , by=.(LinkId)]
merge_hb[, c("hb_flagClosest") := ifelse(hb_diffFromDiag == min(hb_diffFromDiag), 1, 0) , by=.(LinkId)]
merge_hb <- merge_hb[(hb_diffFromDiag < paramFromDiagnosisWindowSeconds) & hb_flagClosest == 1]
merge_hb <- merge_hb[diff(merge_hb$LinkId) != 0]
# SBP
merge_sbp <- merge(merge_hb, sbp_DT_forMerge, by = "LinkId")
merge_sbp[, c("sbp_diffFromDiag") := sqrt((sbp_dateplustime1 - diagnosisDate_unix) ^ 2) , by=.(LinkId)]
merge_sbp[, c("sbp_flagClosest") := ifelse(sbp_diffFromDiag == min(sbp_diffFromDiag), 1, 0) , by=.(LinkId)]
merge_sbp <- merge_sbp[(sbp_diffFromDiag < paramFromDiagnosisWindowSeconds) & sbp_flagClosest == 1]
merge_sbp <- merge_sbp[diff(merge_sbp$LinkId) != 0]
# DBP
merge_dbp <- merge(merge_sbp, dbp_DT_forMerge, by = "LinkId")
merge_dbp[, c("dbp_diffFromDiag") := sqrt((dbp_dateplustime1 - diagnosisDate_unix) ^ 2) , by=.(LinkId)]
merge_dbp[, c("dbp_flagClosest") := ifelse(dbp_diffFromDiag == min(dbp_diffFromDiag), 1, 0) , by=.(LinkId)]
merge_dbp <- merge_dbp[(dbp_diffFromDiag < paramFromDiagnosisWindowSeconds) & dbp_flagClosest == 1]
merge_dbp <- merge_dbp[diff(merge_dbp$LinkId) != 0]
# BMI
merge_bmi <- merge(merge_dbp, bmi_DT_forMerge, by = "LinkId")
merge_bmi[, c("bmi_diffFromDiag") := sqrt((bmi_dateplustime1 - diagnosisDate_unix) ^ 2) , by=.(LinkId)]
merge_bmi[, c("bmi_flagClosest") := ifelse(bmi_diffFromDiag == min(bmi_diffFromDiag), 1, 0) , by=.(LinkId)]
merge_bmi <- merge_bmi[(bmi_diffFromDiag < paramFromDiagnosisWindowSeconds) & bmi_flagClosest == 1]
merge_bmi <- merge_bmi[diff(merge_bmi$LinkId) != 0]
# renal
# merge_renal <- merge(merge_bmi, renal_DT_forMerge, by = "LinkId")
# merge_renal[, c("egfr_diffFromDiag") := sqrt((egfr_dateplustime1 - diagnosisDate_unix) ^ 2) , by=.(LinkId)]
# merge_renal[, c("egfr_flagClosest") := ifelse(egfr_diffFromDiag == min(egfr_diffFromDiag), 1, 0) , by=.(LinkId)]
#
# merge_renal <- merge_renal[(egfr_diffFromDiag < paramFromDiagnosisWindowSeconds) & egfr_flagClosest == 1]
# finalset
diagnostic_test_set <- merge_bmi
# remove duplicates
#diagnostic_test_set <- unique(diagnostic_test_set)
diagnostic_test_set$Sex <- ifelse(diagnostic_test_set$Sex == "Male", 1, 0)
diagnostic_test_set$diabetesType <- ifelse(diagnostic_test_set$diabetesType == "Type 1 Diabetes Mellitus", 1, 0)
diagnostic_test_set <- diagnostic_test_set[substr(Ethnicity,1,3) != "ECD"]
# table(diagnostic_test_set$Ethnicity)
# table(subset(diagnostic_test_set, diabetesType == 1)$Ethnicity)
# table(subset(diagnostic_test_set, diabetesType == 0)$Ethnicity)
factorEthnicity <- factor(diagnostic_test_set$Ethnicity)
diagnostic_test_set$Ethnicity <- as.numeric(factorEthnicity)
diagnostic_test_set$ageAtDiagnosis <- (diagnostic_test_set$diagnosisDate_unix - diagnostic_test_set$DOB_unix) / (60*60*24*365.25)
diagnostic_test_set_withID <- data.table(diagnostic_test_set$LinkId, diagnostic_test_set$ageAtDiagnosis, diagnostic_test_set$Ethnicity, diagnostic_test_set$Sex, diagnostic_test_set$hba1cNumeric, diagnostic_test_set$sbpNumeric, diagnostic_test_set$dbpNumeric, diagnostic_test_set$bmiNumeric, diagnostic_test_set$diabetesType)
colnames(diagnostic_test_set_withID) <- c("LinkId", "age", "ethnicity", "sex", "hba1c", "sbp", "dbp", "bmi", "diabetesType")
diagnostic_test_set <- data.table(diagnostic_test_set$ageAtDiagnosis, diagnostic_test_set$Ethnicity, diagnostic_test_set$Sex, diagnostic_test_set$hba1cNumeric, diagnostic_test_set$sbpNumeric, diagnostic_test_set$dbpNumeric, diagnostic_test_set$bmiNumeric, diagnostic_test_set$diabetesType)
colnames(diagnostic_test_set) <- c("age", "ethnicity", "sex", "hba1c", "sbp", "dbp", "bmi", "diabetesType")
# summary(diagnostic_test_set)
# summary(subset(diagnostic_test_set, diabetesType == 1))
# summary(subset(diagnostic_test_set, diabetesType == 0))
write.table(diagnostic_test_set, file = "~/R/_workingDirectory/t1_t2_ANN/diagSet_7p.csv", sep = ",", row.names = FALSE, col.names = TRUE)
write.table(diagnostic_test_set_withID, file = "~/R/_workingDirectory/t1_t2_ANN/diagSet_7p_withID.csv", sep = ",", row.names = FALSE, col.names = TRUE)
## read in csv output from ann
## turn ethnicity levels into readable output for physician interpretation
levelKey <- levels(factorEthnicity)
X_test_RN <- read.csv("~/R/_workingDirectory/t1_t2_ANN/output/X_test_realNumbers.csv", sep=",", header = FALSE, row.names = NULL)
y_test <- read.csv("~/R/_workingDirectory/t1_t2_ANN/output/y_test.csv", sep=",", header = FALSE, row.names = NULL)
y_pred <- read.csv("~/R/_workingDirectory/t1_t2_ANN/output/y_pred.csv", sep=",", header = FALSE, row.names = NULL)
library(ROCR)
pred <- prediction(y_pred, y_test)
roc.perf <- performance(pred, measure = 'tpr', x.measure = 'fpr')
plot(roc.perf)
# find sensitvity / specifictiy / optimal cut point
opt.cut = function(perf, pred){
cut.ind = mapply(FUN=function(x, y, p){
d = (x - 0)^2 + (y-1)^2
ind = which(d == min(d))
c(sensitivity = y[[ind]], specificity = 1-x[[ind]],
cutoff = p[[ind]])
}, perf@x.values, perf@y.values, pred@cutoffs)
}
print(opt.cut(roc.perf, pred))
X_test_RN$ethnicity_decoded <- ""
levelKey_dummyVarTrapAdj <- levelKey[2: length(levelKey)]
for (r in seq(1, nrow(X_test_RN), 1)) {
eth_sub <- X_test_RN[r, ][1:24]
if (sum(eth_sub) == 0) {
X_test_RN$ethnicity_decoded[r] = "nil recorded"
}
if (sum(eth_sub) == 1) {
ethnicityString <- levelKey_dummyVarTrapAdj[eth_sub == 1]
}
X_test_RN$ethnicity_decoded[r] <- ethnicityString
}
set.seed(42)
# cut redundant ethnicity keys
X_test_RN_cut <- X_test_RN[25: ncol(X_test_RN)]
colnames(X_test_RN_cut) <- c("age", "male", "hba1c", "sbp", "dbp", "bmi", "ethnicity")
X_test_RN_withTrueType <- cbind(X_test_RN_cut, y_test)
colnames(X_test_RN_withTrueType) <- c("age", "male", "hba1c", "sbp", "dbp", "bmi", "ethnicity", "DMtype")
X_test_RN_withTrueType_withProb <- cbind(X_test_RN_withTrueType, y_pred)
colnames(X_test_RN_withTrueType_withProb) <- c("age", "male", "hba1c", "sbp", "dbp", "bmi", "ethnicity", "DMtype", "DMpred")
X_test_RN_withTrueType_withProbDT <- data.table(X_test_RN_withTrueType_withProb)
t1_table <- X_test_RN_withTrueType_withProbDT[DMtype == 1]
t2_table <- X_test_RN_withTrueType_withProbDT[DMtype == 0]
random_t2_sample <- t2_table[sample(nrow(t2_table), (nrow(t1_table) * 3)), ]
physician_pool <- rbind(t1_table, random_t2_sample)
# first set - for greg
physician_set1 <- physician_pool[sample(nrow(physician_pool), 100), ]
physician_set1_write1 <- subset(physician_set1, select = c(1:7))
write.table(physician_set1, file = "~/R/_workingDirectory/t1_t2_ANN/output/physician_set1_key.csv", sep = ",", row.names = FALSE)
write.table(physician_set1_write1, file = "~/R/_workingDirectory/t1_t2_ANN/output/physician_set1.csv", sep = ",", row.names = FALSE)
# automated sets with proportion randomisation
# n per sample
sample_n <- 100
# number of sets
set_n <- 40
for (ii in seq(1, set_n, 1)) {
random_t2_multiplier <- runif(1, 1, 10)
# print(random_t2_multiplier)
random_t2_sample <- t2_table[sample(nrow(t2_table), (nrow(t1_table) * random_t2_multiplier)), ]
physician_pool <- rbind(t1_table, random_t2_sample)
physician_set <- physician_pool[sample(nrow(physician_pool), sample_n), ]
physician_set_write <- subset(physician_set, select = c(1:7))
keyName <- paste('~/R/_workingDirectory/t1_t2_ANN/output/samplesForTesting/physician_set_', ii, '_key.csv', sep = '')
write.table(physician_set, file = keyName, sep = ",", row.names = FALSE)
testSetName <- paste('~/R/_workingDirectory/t1_t2_ANN/output/samplesForTesting/physician_set_', ii, '.csv', sep = '')
write.table(physician_set_write, file = testSetName, sep = ",", row.names = FALSE)
}
## analyse performance
library(caret)
library(ROCR)
performanceAnalysis <- function(physicianPrediction, ANNprediction, key, ANNthreshold) {
ann_pred <- ifelse(ANNprediction > ANNthreshold, 1, 0)
print(confusionMatrix(data = physicianPrediction, reference = key))
cm_phys <- table(key, physicianPrediction); print(cm_phys)
accuracy_phys <- (cm_phys[1,1] + cm_phys[2,2]) / sum(cm_phys); print(accuracy_phys)
print(confusionMatrix(data = ann_pred, reference = key))
cm_ann <- table(key, ann_pred); print(cm_ann)
accuracy_ann <- (cm_ann[1,1] + cm_ann[2,2]) / sum(cm_ann); print(accuracy_ann)
pred <- prediction(ANNprediction, key)
roc.perf <- performance(pred, measure = 'tpr', x.measure = 'fpr')
plot(roc.perf)
auc.perf = performance(pred, measure = "auc")
auc.perf@y.values
print(auc.perf)
points((cm_phys[2, 1] / sum(cm_phys[2, ])), (cm_phys[2, 2] / sum(cm_phys[2, ])), col = "red", pch = 16, cex = 2)
points((cm_ann[2, 1] / sum(cm_ann[2, ])), (cm_ann[2, 2] / sum(cm_ann[2, ])), col = "blue", pch = 16, cex = 2)
}
perf_1 <- read.csv('~/R/_workingDirectory/t1_t2_ANN/output/samplesForTesting/physician_set_1_done.csv', header = T)
perf_1_key <- read.csv('~/R/_workingDirectory/t1_t2_ANN/output/samplesForTesting/physician_set_1_key.csv', header = T)
performanceAnalysis(perf_1$diagnosis, perf_1_key$DMpred, perf_1_key$DMtype, 0.05)
perf_1 <- read.csv('~/R/_workingDirectory/t1_t2_ANN/output/samplesForTesting/physician_set_38_CS_done.csv', header = T)
perf_1_key <- read.csv('~/R/_workingDirectory/t1_t2_ANN/output/samplesForTesting/physician_set_38_key.csv', header = T)
performanceAnalysis(perf_1$diagnosis, perf_1_key$DMpred, perf_1_key$DMtype, 0.05)
## analyse performance
perf_1 <- read.csv('~/R/_workingDirectory/t1_t2_ANN/output/physician_set1_GJ130617.csv', header = T)
perf_1_key <- read.csv('~/R/_workingDirectory/t1_t2_ANN/output/physician_set1_key.csv', header = T)
performanceAnalysis(perf_1$diagnosis, perf_1_key$DMpred, perf_1_key$DMtype, 0.05)
## reconstruct original ID list
###############################################################################################
## 140717 run from here
diagnostic_test_set_withID <- read.csv("~/R/_workingDirectory/t1_t2_ANN/diagSet_7p_withID.csv")
nameList_physicianResults <- c('physician_set_1_done', 'physician_set_6', 'physician_set_11', 'physician_set_16', 'physician_set_22', 'physician_set_38', 'physician_set_23', 'physician_set_24')
nameList_keys <- c('physician_set_1_key', 'physician_set_6_key', 'physician_set_11_key', 'physician_set_16_key', 'physician_set_22_key', 'physician_set_38_key', 'physician_set_23_key', 'physician_set_24_key')
for (k in seq(1, length(nameList_physicianResults), 1)) {
phys_name <- paste('~/R/_workingDirectory/t1_t2_ANN/output/completedSets/',nameList_physicianResults[k], '.csv', sep = '')
key_name <- paste('~/R/_workingDirectory/t1_t2_ANN/output/completedSets/',nameList_keys[k], '.csv', sep = '')
physicianSet <- read.table(phys_name, header = T, row.names = NULL, sep = ',')
colnames(physicianSet) <- c('age', 'male', 'hba1c', 'sbp', 'dbp', 'bmi', 'ethnicity', 'diagnosis')
keySet <- read.table(key_name, header = T, row.names = NULL, sep = ',')
keySet$physicianDiagnosis <- physicianSet$diagnosis
set_includingIDs <- merge(keySet, as.data.frame(diagnostic_test_set_withID), by.x = c("hba1c", "sbp", "dbp", "bmi"), by.y = c("hba1c", "sbp", "dbp", "bmi"))
set_includingIDs <- unique(set_includingIDs)
# if running individual file, use this statement
# outputFrame <- set_includingIDs
if (k == 1) { outputFrame <- set_includingIDs }
if (k > 1) { outputFrame <- rbind(outputFrame, set_includingIDs) }
}
write.table(outputFrame, file = "~/R/_workingDirectory/t1_t2_ANN/outputFrame.csv", sep = ",", row.names = FALSE, col.names = TRUE)
# performanceAnalysis(outputFrame$physicianDiagnosis, outputFrame$DMpred, outputFrame$DMtype, 0.04306091)
performanceAnalysis(outputFrame$physicianDiagnosis, outputFrame$DMpred, outputFrame$DMtype, 0.1)
########### run r_logit_model.R and come back
logitInput <- read.csv("~/R/_workingDirectory/t1_t2_ANN/logit_output.csv")
logitInput <- data.table(logitInput)
logitInput = unique(logitInput)
outputFrame <- data.table(outputFrame)
outputFrame$prob_pred <- (-1)
for (j in seq(1, nrow(outputFrame), 1)) {
idOfInterest <- outputFrame$LinkId[j]
logitSub <- logitInput[LinkId == idOfInterest]
if(nrow(logitSub == 1)) {
outputFrame$prob_pred[j] <- logitSub$prob_pred
}
}
# performanceAnalysis(outputFrame$physicianDiagnosis, outputFrame$prob_pred, outputFrame$DMtype, 0.06596188)
performanceAnalysis(outputFrame$physicianDiagnosis, outputFrame$prob_pred, outputFrame$DMtype, 0.1)
outputFrame$ann_binary <- ifelse(outputFrame$DMpred > 0.1, 1, 0)
outputFrame$logit_binary <- ifelse(outputFrame$prob_pred > 0.1, 1, 0)
physicianConcordance_ann <- sum(ifelse((outputFrame$ann_binary == 1 & outputFrame$physicianDiagnosis == 1) | (outputFrame$ann_binary == 0 & outputFrame$physicianDiagnosis == 0), 1, 0)) / nrow(outputFrame)
print(physicianConcordance_ann)
physicianConcordance_logit <- sum(ifelse((outputFrame$logit_binary == 1 & outputFrame$physicianDiagnosis == 1) | (outputFrame$logit_binary == 0 & outputFrame$physicianDiagnosis == 0), 1, 0)) / nrow(outputFrame)
print(physicianConcordance_logit)
ann_concordance_logit <- sum(ifelse((outputFrame$ann_binary == 1 & outputFrame$logit_binary == 1) | (outputFrame$ann_binary == 0 & outputFrame$logit_binary == 0), 1, 0)) / nrow(outputFrame)
print(ann_concordance_logit)
# flag those where ann and logit agree, but diagnosis different
high_prob_misdiagnosis <- outputFrame[(ann_binary == logit_binary) & (ann_binary != DMtype)]
# lm probabilities / residuals
plot(outputFrame$DMpred, outputFrame$prob_pred)
regressionLine <- lm(outputFrame$prob_pred ~ outputFrame$DMpred)
abline(regressionLine, col = 'red')
lm_residuals <- resid(regressionLine)
outputFrame$residuals <- lm_residuals
outputFrame_orderedByResiduals <- outputFrame[order(outputFrame$residuals), ]
# plot relations
plot(outputFrame$bmi, outputFrame$DMpred)
plot(outputFrame$bmi, outputFrame$prob_pred)
plot(outputFrame$age.x, outputFrame$DMpred)
ann_age_boxplot <- boxplot(outputFrame$DMpred ~ cut(outputFrame$age.x, breaks = seq(0, 100, 5)))
logit_age_boxplot <- boxplot(outputFrame$prob_pred ~ cut(outputFrame$age.x, breaks = seq(0, 100, 5)))
plot(ann_age_boxplot$stats[3, ]); lines(ann_age_boxplot$stats[3, ])
points(logit_age_boxplot$stats[3, ], col = 'red'); lines(logit_age_boxplot$stats[3, ], col = 'red')
ann_bmi_boxplot <- boxplot(outputFrame$DMpred ~ cut(outputFrame$bmi, breaks = seq(10, 50, 2)))
logit_bmi_boxplot <- boxplot(outputFrame$prob_pred ~ cut(outputFrame$bmi, breaks = seq(10, 50, 2)))
plot(ann_bmi_boxplot$stats[3, ]); lines(ann_bmi_boxplot$stats[3, ])
points(logit_bmi_boxplot$stats[3, ], col = 'red'); lines(logit_bmi_boxplot$stats[3, ], col = 'red')
ann_hba1c_boxplot <- boxplot(outputFrame$DMpred ~ cut(outputFrame$hba1c, breaks = seq(10, 100, 5)))
logit_hba1c_boxplot <- boxplot(outputFrame$prob_pred ~ cut(outputFrame$hba1c, breaks = seq(10, 100, 5)))
plot(ann_hba1c_boxplot$stats[3, ]); lines(ann_hba1c_boxplot$stats[3, ])
points(logit_hba1c_boxplot$stats[3, ], col = 'red'); lines(logit_hba1c_boxplot$stats[3, ], col = 'red')
ann_sbp_boxplot <- boxplot(outputFrame$DMpred ~ cut(outputFrame$sbp, breaks = seq(80, 200, 10)))
logit_sbp_boxplot <- boxplot(outputFrame$prob_pred ~ cut(outputFrame$sbp, breaks = seq(80, 200, 10)))
plot(ann_sbp_boxplot$stats[3, ]); lines(ann_sbp_boxplot$stats[3, ])
points(logit_sbp_boxplot$stats[3, ], col = 'red'); lines(logit_sbp_boxplot$stats[3, ], col = 'red')
plot(outputFrame$age.x, outputFrame$prob_pred)
plot(outputFrame$age.x, outputFrame$DMpred)
plot(outputFrame$sbp, outputFrame$DMpred)
plot(outputFrame$dbp, outputFrame$DMpred)
boxplot(outputFrame$prob_pred ~ outputFrame$ethnicity.x, las=3, varwidth = T)
boxplot(outputFrame$DMpred ~ outputFrame$ethnicity.x, las=3, varwidth = T)
plot(outputFrame$hba1c, outputFrame$DMpred)
plot(outputFrame$hba1c, outputFrame$prob_pred)
plot(outputFrame$hba1c, outputFrame$DMpred)
points(outputFrame$hba1c, outputFrame$prob_pred, col = 'red')