Skip to content

Latest commit

 

History

History
82 lines (61 loc) · 2.37 KB

README.md

File metadata and controls

82 lines (61 loc) · 2.37 KB

Apriori-Algorithm

Frequent Itemset Generation and Association Rule Mining

Includes:

  • Apriori Algorithm: To find frequent-itemsets from a set of transactions.
  • Rule generation: Generation of 'interesting' rules from the frequent itemset.
  • Hashing: Itemsets are hashed to get their support count in almost constant time.

Requirements:

Only python3 is required to run this algorithm. No need to install anything else.

python3 is installed in most Linux distributions, by default.

Mining the Rules:

First make the Apriori file executable and then simply run it as:

$ chmod +x ./Apriori
$ ./Apriori
This will make a directory: results at your current path and store the results in the in a file.

Name of file: It will be the values of support and confidence for which Apriori Algorithm is run.

Paremeters:

Support: Defaults value of support is 0.02

Confidence: Default value of confidence is 0.45

Input Transaction file: Currently set as groceries.csv.

All the parameters can be changed from within the code.

Output:

In a file s=0.01 c=0.5

Frequnet Itemsets:
1-itemsets:
['liquor'] | support: 109
['dessert'] | support: 365
['sliced cheese'] | support: 241
['bottled water'] | support: 1087
['oil'] | support: 276
['yogurt'] | support: 1372
....
Count: 88

2-itemsets:
['whole milk', 'citrus fruit'] | support: 300
['other vegetables', 'margarine'] | support: 194
['whipped/sour cream', 'citrus fruit'] | support: 107
['whole milk', 'cream cheese'] | support: 162
['rolls/buns', 'citrus fruit'] | support: 165
['soda', 'citrus fruit'] | support: 126
...
Count: 213

...
Total number of frequent itemset(s): 333


Rules:
['citrus fruit', 'root vegetables'](174) -> ['other vegetables'](1903) | confidence:
0.5862068965517241
['root vegetables', 'tropical fruit'](207) -> ['other vegetables'](1903) | confidence:
0.5845410628019324
['curd', 'yogurt'](170) -> ['whole milk'](2513) | confidence: 0.5823529411764706
['other vegetables', 'butter'](197) -> ['whole milk'](2513) | confidence:
0.5736040609137056
['root vegetables', 'tropical fruit'](207) -> ['whole milk'](2513) | confidence:
0.5700483091787439
....
Total number of rules: 14

Format of Rules:

[LHS] (item set (count)) -> [RHS] (item set (count)) | confidence: confidence value