2139472 Lab_3

December 1, 2021

1 Lab Program 3

2 Vinay Sirohi

3 2139472
3.1 Identify frequent item sets using Apriori Algorithm
[54]: #Importing mlxtend and printing the current version of it.
import mlxtend
print(mlxtend.__version__)

0.19.0

[2]: #Importing important libraries of our use

from mlxtend.frequent_patterns import apriori,association_rules
import numpy as np
import pandas as pd
from [Link] import TransactionEncoder
import time

[12]: #Reading out data set

df = pd.read_csv('Groceries_dataset.csv')
#Looking at the top 20 values of it.
[Link](20)

[12]: Member_number Date itemDescription

0 1808 21-07-2015 tropical fruit
1 2552 05-01-2015 whole milk
2 2300 19-09-2015 pip fruit
3 1187 12-12-2015 other vegetables
4 3037 01-02-2015 whole milk
5 4941 14-02-2015 rolls/buns
6 4501 08-05-2015 other vegetables
7 3803 23-12-2015 pot plants
8 2762 20-03-2015 whole milk
9 4119 12-02-2015 tropical fruit
10 1340 24-02-2015 citrus fruit

1
 11 2193 14-04-2015 beef
12 1997 21-07-2015 frankfurter
13 4546 03-09-2015 chicken
14 4736 21-07-2015 butter
15 1959 30-03-2015 fruit/vegetable juice
16 1974 03-05-2015 packaged fruit/vegetables
17 2421 02-09-2015 chocolate
18 1513 03-08-2015 specialty bar
19 1905 07-07-2015 other vegetables

[24]: #Converting date column to proper format

df['Date'] = pd.to_datetime(df['Date'])
df

[24]: Member_number Date itemDescription

0 1808 2015-07-21 tropical fruit
1 2552 2015-05-01 whole milk
2 2300 2015-09-19 pip fruit
3 1187 2015-12-12 other vegetables
4 3037 2015-01-02 whole milk
… … … …
38760 4471 2014-08-10 sliced cheese
38761 2022 2014-02-23 candy
38762 1097 2014-04-16 cake bar
38763 1510 2014-03-12 fruit/vegetable juice
38764 1521 2014-12-26 cat food

[38765 rows x 3 columns]

[27]: #Grouping the transactions by member number and date

df['itemDescription'] = [Link](['Member_number',␣
,→'Date'])['itemDescription'].transform(lambda x: ','.join(x))

df

[27]: Member_number Date \

0 1808 2015-07-21
1 2552 2015-05-01
2 2300 2015-09-19
3 1187 2015-12-12
4 3037 2015-01-02
… … …
38760 4471 2014-08-10
38761 2022 2014-02-23
38762 1097 2014-04-16
38763 1510 2014-03-12
38764 1521 2014-12-26

2
 itemDescription
0 tropical fruit,rolls/buns,candy,tropical fruit…
1 whole milk,tropical fruit,chocolate,whole milk…
2 pip fruit,other vegetables,flour,pip fruit,oth…
3 other vegetables,onions,shopping bags,other ve…
4 whole milk,other vegetables,white bread,whole …
… …
38760 whole milk,yogurt,sliced cheese,whole milk,yog…
38761 cat food,yogurt,candy,cat food,yogurt,candy,ca…
38762 sausage,whole milk,cake bar,sausage,whole milk…
38763 beef,canned beer,fruit/vegetable juice,beef,ca…
38764 ham,seasonal products,cat food,ham,seasonal pr…

[38765 rows x 3 columns]

[34]: #Let's now create a list of the transactions so that we can transform our data␣
,→into the correct format using TransactionEncoder.

df1=[]
for i in range(0,len(df)-1):
data = df['itemDescription'][i].split(',')
[Link](data)

[59]: #Applying transaction encoder to our data set

te = TransactionEncoder()
te_ary = [Link](df1).transform(df1)
df2 = [Link](te_ary, columns=te.columns_)
df2[:10]

[59]: Instant food products UHT-milk abrasive cleaner artif. sweetener \

0 False False False False
1 False False False False
2 False False False False
3 False False False False
4 False False False False
5 False False False False
6 False False False False
7 False False False False
8 False False False False
9 False False False False

baby cosmetics bags baking powder bathroom cleaner beef berries \

0 False False False False False False
1 False False False False False False
2 False False False False False False
3 False False False False False False
4 False False False False False False
5 False False False False False False

3
 6 False False False False False False
7 False False False False False False
8 False False False False False False
9 False False False False False False

… turkey vinegar waffles whipped/sour cream whisky white bread \

0 … False False False False False False
1 … False False False False False False
2 … False False False False False False
3 … False False False False False False
4 … False False False False False True
5 … False False False False False False
6 … False False False False False False
7 … False False False True False False
8 … False False False False False False
9 … False False False False False False

white wine whole milk yogurt zwieback

0 False False False False
1 False True False False
2 False False False False
3 False False False False
4 False True False False
5 False False False False
6 False False False False
7 False False False False
8 False True False False
9 False False False False

[10 rows x 167 columns]

[62]: #Applying apriori to look at itemsets which have a minimum support level of 0.02
frq = apriori(df2, min_support=0.02)
[Link]()

[62]: support itemsets

0 0.024533 (1)
1 0.039624 (8)
2 0.024739 (9)
3 0.054742 (11)
4 0.069497 (12)

A float between 0 and 1 for minimum support of the itemsets returned. The support is computed
as the fraction transactions_where_item(s)_occur / total_transactions.

[63]: #Use colnames = True to look at the itemsets by name

fr = apriori(df2, min_support=0.02, use_colnames=True)
[Link]()

4
[63]: support itemsets
0 0.024533 (UHT-milk)
1 0.039624 (beef)
2 0.024739 (berries)
3 0.054742 (bottled beer)
4 0.069497 (bottled water)

[42]: #applying lambda function to look at the length of the itemsets

frequent_itemsets = apriori(df2, min_support=0.02, use_colnames=True)
frequent_itemsets['length'] = frequent_itemsets['itemsets'].apply(lambda x:␣
,→len(x))

frequent_itemsets

[42]: support itemsets length

0 0.024533 (UHT-milk) 1
1 0.039624 (beef) 1
2 0.024739 (berries) 1
3 0.054742 (bottled beer) 1
4 0.069497 (bottled water) 1
5 0.044165 (brown bread) 1
6 0.041895 (butter) 1
7 0.020586 (butter milk) 1
8 0.055954 (canned beer) 1
9 0.033072 (chicken) 1
10 0.028093 (chocolate) 1
11 0.060649 (citrus fruit) 1
12 0.036658 (coffee) 1
13 0.027035 (cream cheese ) 1
14 0.040914 (curd) 1
15 0.026029 (dessert) 1
16 0.044139 (domestic eggs) 1
17 0.047983 (frankfurter) 1
18 0.020638 (frozen meals) 1
19 0.033201 (frozen vegetables) 1
20 0.040140 (fruit/vegetable juice) 1
21 0.025384 (hamburger meat) 1
22 0.021128 (long life bakery product) 1
23 0.038464 (margarine) 1
24 0.026287 (napkins) 1
25 0.046048 (newspapers) 1
26 0.023140 (onions) 1
27 0.137679 (other vegetables) 1
28 0.060623 (pastry) 1
29 0.057708 (pip fruit) 1
30 0.043778 (pork) 1
31 0.127954 (rolls/buns) 1
32 0.079404 (root vegetables) 1

5
 33 0.022289 (salty snack) 1
34 0.074580 (sausage) 1
35 0.056444 (shopping bags) 1
36 0.112269 (soda) 1
37 0.021128 (sugar) 1
38 0.079326 (tropical fruit) 1
39 0.022315 (waffles) 1
40 0.048344 (whipped/sour cream) 1
41 0.028841 (white bread) 1
42 0.183753 (whole milk) 1
43 0.101434 (yogurt) 1
44 0.022340 (other vegetables, whole milk) 2
45 0.022237 (whole milk, rolls/buns) 2

[48]: #selecting itemsets which have a minimum support level of 0.02 and lenght 2
frequent_itemsets[ (frequent_itemsets['length'] == 2) &
(frequent_itemsets['support'] > 0.02) ]

[48]: support itemsets length

44 0.022340 (other vegetables, whole milk) 2
45 0.022237 (whole milk, rolls/buns) 2

[61]: #Applying association rules with minimum threshold value 0.02

rules = association_rules(frequent_itemsets , metric="confidence",␣
,→min_threshold=0.02)

rules

[61]: antecedents consequents antecedent support \

0 (other vegetables) (whole milk) 0.137679
1 (whole milk) (other vegetables) 0.183753
2 (whole milk) (rolls/buns) 0.183753
3 (rolls/buns) (whole milk) 0.127954

consequent support support confidence lift leverage conviction

0 0.183753 0.022340 0.162263 0.883052 -0.002959 0.974348
1 0.137679 0.022340 0.121578 0.883052 -0.002959 0.981670
2 0.127954 0.022237 0.121016 0.945782 -0.001275 0.992108
3 0.183753 0.022237 0.173790 0.945782 -0.001275 0.987942