Data Mining - Association Rules

import pandas as pd import sidetable from mlxtend.preprocessing import TransactionEncoder from mlxtend.frequent_patterns import apriori, fpgrowth

groceries_df = pd.read_csv('https://gist.githubusercontent.com/ahbanavi/0b1f62713bdf8bd32da0e17ae07e0fad/raw/3974126661c5d25505ead0b23dee70eee2a4f83a/Groceries_dataset.csv') groceries_df.head()

Member_numberint64

Dateobject

1808

21-07-2015

2552

05-01-2015

2300

19-09-2015

1187

12-12-2015

3037

01-02-2015

groceries_df.rename(columns={'Member_number': 'customer_id', 'itemDescription': 'item'},inplace=True) groceries_df.drop(columns=['Date'], inplace=True) print("Items count before deleting duplicates: ", groceries_df.shape[0]) groceries_df.drop_duplicates(inplace=True) print("Items count after deleting duplicates: ", groceries_df.shape[0]) display(groceries_df)

Items count before deleting duplicates:  38765
Items count after deleting duplicates:  34766

customer_idint64

1000 - 5000

itemobject

whole milk5.1%

other vegetables4.2%

165 others90.6%

1808

tropical fruit

2552

whole milk

2300

pip fruit

1187

other vegetables

3037

whole milk

4941

rolls/buns

4501

other vegetables

3803

pot plants

2762

whole milk

4119

tropical fruit

items_set = groceries_df.groupby(by = ['customer_id'])['item'].apply(list).reset_index() display(items_set)

customer_idint64

1000 - 5000

itemobject

['sausage', 'soda']0.1%

['frankfurter', 'other vegetables']0.1%

3878 others99.8%

1000

['soda', 'canned beer', 'sausage', 'whole milk', 'pickled vegetables', 'misc. beverages', 'semi-fin…

1001

['frankfurter', 'beef', 'sausage', 'whole milk', 'soda', 'curd', 'white bread', 'whipped/sour cream…

1002

['tropical fruit', 'butter milk', 'butter', 'frozen vegetables', 'sugar', 'specialty chocolate', 'w…

1003

['sausage', 'root vegetables', 'rolls/buns', 'detergent', 'frozen meals', 'dental care']

1004

['other vegetables', 'pip fruit', 'root vegetables', 'canned beer', 'rolls/buns', 'whole milk', 'hy…

1005

['whipped/sour cream', 'rolls/buns', 'margarine']

1006

['whole milk', 'frankfurter', 'chicken', 'bottled water', 'flour', 'chocolate', 'bottled beer', 'ro…

1008

['hamburger meat', 'tropical fruit', 'soda', 'liquor (appetizer)', 'yogurt', 'liver loaf', 'root ve…

1009

['herbs', 'pastry', 'tropical fruit', 'yogurt', 'ketchup', 'canned fish', 'newspapers', 'cocoa drin…

1010

['pip fruit', 'frankfurter', 'bottled water', 'candles', 'coffee', 'specialty bar', 'kitchen towels…

items_list = items_set['item'].tolist() te = TransactionEncoder() te_ary = te.fit_transform(items_list) items_df = pd.DataFrame(te_ary, columns=te.columns_) display(items_df)

Instant food productsbool

False98.5%

True1.5%

UHT-milkbool

False92.1%

True7.9%

false

true

frequent_itemsets = apriori(items_df, min_support=0.01, use_colnames=True) # Add length of itemsets to dataframe frequent_itemsets['length'] = frequent_itemsets['itemsets'].apply(lambda x: len(x)) frequent_itemsets.sort_values(by='support', inplace=True, ascending=False) display(frequent_itemsets)

supportfloat64

0.010005130836326322 - 0.4581836839404823

itemsetsobject

frozenset({'whole milk'})0%

frozenset({'other vegetables'})0%

3014 others99.9%

113

0.4581836839

frozenset({'whole milk'})

0.3766033864

frozenset({'other vegetables'})

0.3496664956

frozenset({'rolls/buns'})

0.3134940995

frozenset({'soda'})

114

0.2829656234

frozenset({'yogurt'})

106

0.2337095947

frozenset({'tropical fruit'})

0.2306310929

frozenset({'root vegetables'})

0.213699333

frozenset({'bottled water'})

0.2060030785

frozenset({'sausage'})

1050

0.191380195

frozenset({'whole milk', 'other vegetables'})

def display_itemset(n): display(str(n) + " Members Frequent Itemsets", frequent_itemsets[frequent_itemsets['length'] == n])

display_itemset(2)

supportfloat64

0.010005130836326322 - 0.1913801949717804

itemsetsobject

frozenset({'whole milk', 'other vegetables'})0.1%

frozenset({'whole milk', 'rolls/buns'})0.1%

1124 others99.8%

1050

0.191380195

frozenset({'whole milk', 'other vegetables'})

1140

0.1785531042

frozenset({'whole milk', 'rolls/buns'})

1209

0.1511031298

frozenset({'whole milk', 'soda'})

1241

0.1505900462

frozenset({'whole milk', 'yogurt'})

1031

0.1467419189

frozenset({'other vegetables', 'rolls/buns'})

1039

0.1241662391

frozenset({'other vegetables', 'soda'})

1051

0.1203181119

frozenset({'other vegetables', 'yogurt'})

1130

0.1198050282

frozenset({'rolls/buns', 'soda'})

1229

0.1164699846

frozenset({'whole milk', 'tropical fruit'})

1157

0.113134941

frozenset({'whole milk', 'root vegetables'})

display_itemset(3)

supportfloat64

0.010005130836326322 - 0.08209338122113904

itemsetsobject

frozenset({'whole milk', 'other vegetables', 'rolls/buns'})0.1%

frozenset({'whole milk', 'other vegetables', 'yogurt'})0.1%

1457 others99.9%

2447

0.08209338122

frozenset({'whole milk', 'other vegetables', 'rolls/buns'})

2496

0.07183170857

frozenset({'whole milk', 'other vegetables', 'yogurt'})

2480

0.06926629041

frozenset({'whole milk', 'other vegetables', 'soda'})

2626

0.06593124679

frozenset({'whole milk', 'yogurt', 'rolls/buns'})

2611

0.06516162134

frozenset({'whole milk', 'rolls/buns', 'soda'})

1532

0.05618265777

frozenset({'whole milk', 'other vegetables', 'bottled water'})

2688

0.05438686506

frozenset({'whole milk', 'yogurt', 'soda'})

2439

0.05259107234

frozenset({'other vegetables', 'rolls/buns', 'soda'})

2448

0.05233453053

frozenset({'yogurt', 'other vegetables', 'rolls/buns'})

2487

0.05053873781

frozenset({'whole milk', 'other vegetables', 'tropical fruit'})

display_itemset(4)

supportfloat64

0.010005130836326322 - 0.03437660338635198

itemsetsobject

frozenset({'whole milk', 'yogurt', 'other vegetables', 'rolls/buns'})0.3%

frozenset({'whole milk', 'other vegetables', 'rolls/buns', 'soda'})0.3%

309 others99.4%

2765

0.01128783992

frozenset({'whole milk', 'shopping bags', 'bottled water', 'rolls/buns'})

2808

0.01128783992

frozenset({'whole milk', 'canned beer', 'tropical fruit', 'rolls/buns'})

2810

0.01128783992

frozenset({'whole milk', 'root vegetables', 'canned beer', 'soda'})

2813

0.01128783992

frozenset({'whole milk', 'chocolate', 'other vegetables', 'yogurt'})

2849

0.01128783992

frozenset({'whole milk', 'other vegetables', 'yogurt', 'domestic eggs'})

2774

0.01128783992

frozenset({'whole milk', 'sausage', 'yogurt', 'bottled water'})

2981

0.01128783992

frozenset({'whole milk', 'shopping bags', 'sausage', 'rolls/buns'})

2775

0.01128783992

frozenset({'whole milk', 'shopping bags', 'bottled water', 'soda'})

2898

0.01128783992

frozenset({'whole milk', 'sausage', 'other vegetables', 'pip fruit'})

2884

0.01128783992

frozenset({'whole milk', 'other vegetables', 'pastry', 'pip fruit'})

display_itemset(5)

supportfloat64

itemsetsobject

3015

0.01359671626

frozenset({'whole milk', 'rolls/buns', 'other vegetables', 'soda', 'yogurt'})

3013

0.01359671626

frozenset({'whole milk', 'rolls/buns', 'sausage', 'other vegetables', 'yogurt'})

3012

0.01051821447

frozenset({'whole milk', 'rolls/buns', 'other vegetables', 'yogurt', 'bottled water'})

3014

0.01000513084

frozenset({'whole milk', 'shopping bags', 'rolls/buns', 'other vegetables', 'yogurt'})

frequent_itemsets = fpgrowth(items_df, min_support=0.01, use_colnames=True) # Add length of itemsets to dataframe frequent_itemsets['length'] = frequent_itemsets['itemsets'].apply(lambda x: len(x)) frequent_itemsets.sort_values(by='support', inplace=True, ascending=False) display(frequent_itemsets)

supportfloat64

0.010005130836326322 - 0.4581836839404823

itemsetsobject

frozenset({'whole milk'})0%

frozenset({'other vegetables'})0%

3014 others99.9%

0.4581836839

frozenset({'whole milk'})

0.3766033864

frozenset({'other vegetables'})

0.3496664956

frozenset({'rolls/buns'})

0.3134940995

frozenset({'soda'})

0.2829656234

frozenset({'yogurt'})

0.2337095947

frozenset({'tropical fruit'})

0.2306310929

frozenset({'root vegetables'})

0.213699333

frozenset({'bottled water'})

0.2060030785

frozenset({'sausage'})

862

0.191380195

frozenset({'whole milk', 'other vegetables'})

display_itemset(2)

supportfloat64

0.010005130836326322 - 0.1913801949717804

itemsetsobject

frozenset({'whole milk', 'other vegetables'})0.1%

frozenset({'whole milk', 'rolls/buns'})0.1%

1124 others99.8%

862

0.191380195

frozenset({'whole milk', 'other vegetables'})

460

0.1785531042

frozenset({'whole milk', 'rolls/buns'})

116

0.1511031298

frozenset({'whole milk', 'soda'})

123

0.1505900462

frozenset({'whole milk', 'yogurt'})

461

0.1467419189

frozenset({'other vegetables', 'rolls/buns'})

118

0.1241662391

frozenset({'other vegetables', 'soda'})

125

0.1203181119

frozenset({'other vegetables', 'yogurt'})

117

0.1198050282

frozenset({'rolls/buns', 'soda'})

863

0.1164699846

frozenset({'whole milk', 'tropical fruit'})

1116

0.113134941

frozenset({'whole milk', 'root vegetables'})

display_itemset(3)

supportfloat64

0.010005130836326322 - 0.08209338122113904

itemsetsobject

frozenset({'whole milk', 'other vegetables', 'rolls/buns'})0.1%

frozenset({'whole milk', 'other vegetables', 'yogurt'})0.1%

1457 others99.9%

462

0.08209338122

frozenset({'whole milk', 'other vegetables', 'rolls/buns'})

134

0.07183170857

frozenset({'whole milk', 'other vegetables', 'yogurt'})

122

0.06926629041

frozenset({'whole milk', 'other vegetables', 'soda'})

135

0.06593124679

frozenset({'whole milk', 'yogurt', 'rolls/buns'})

119

0.06516162134

frozenset({'whole milk', 'rolls/buns', 'soda'})

1522

0.05618265777

frozenset({'whole milk', 'other vegetables', 'bottled water'})

127

0.05438686506

frozenset({'whole milk', 'yogurt', 'soda'})

120

0.05259107234

frozenset({'other vegetables', 'rolls/buns', 'soda'})

136

0.05233453053

frozenset({'yogurt', 'other vegetables', 'rolls/buns'})

868

0.05053873781

frozenset({'whole milk', 'other vegetables', 'tropical fruit'})

display_itemset(4)

supportfloat64

0.010005130836326322 - 0.03437660338635198

itemsetsobject

frozenset({'whole milk', 'yogurt', 'other vegetables', 'rolls/buns'})0.3%

frozenset({'whole milk', 'other vegetables', 'rolls/buns', 'soda'})0.3%

309 others99.4%

137

0.03437660339

frozenset({'whole milk', 'yogurt', 'other vegetables', 'rolls/buns'})

121

0.03181118522

frozenset({'whole milk', 'other vegetables', 'rolls/buns', 'soda'})

130

0.02796305798

frozenset({'whole milk', 'other vegetables', 'yogurt', 'soda'})

165

0.02616726526

frozenset({'whole milk', 'sausage', 'other vegetables', 'rolls/buns'})

132

0.02462801437

frozenset({'whole milk', 'rolls/buns', 'yogurt', 'soda'})

1521

0.02308876347

frozenset({'whole milk', 'bottled water', 'other vegetables', 'rolls/buns'})

162

0.02308876347

frozenset({'whole milk', 'sausage', 'other vegetables', 'yogurt'})

159

0.02283222165

frozenset({'whole milk', 'sausage', 'yogurt', 'rolls/buns'})

1565

0.0220625962

frozenset({'whole milk', 'other vegetables', 'bottled water', 'soda'})

1527

0.0220625962

frozenset({'whole milk', 'yogurt', 'other vegetables', 'bottled water'})

display_itemset(5)

supportfloat64

itemsetsobject

161

0.01359671626

frozenset({'whole milk', 'rolls/buns', 'sausage', 'other vegetables', 'yogurt'})

133

0.01359671626

frozenset({'whole milk', 'rolls/buns', 'other vegetables', 'soda', 'yogurt'})

1530

0.01051821447

frozenset({'whole milk', 'rolls/buns', 'other vegetables', 'yogurt', 'bottled water'})

1327

0.01000513084

frozenset({'whole milk', 'shopping bags', 'rolls/buns', 'other vegetables', 'yogurt'})