Recap on transactions

Análise de Cesta de Compras em R

Christopher Bruffaerts

Statistician

Important points in market basket analysis

Market basket analysis

Focus on the what, not on the how much;
i.e. what do customers have in their baskets?

one_grocery_basket

Main metrics

Support
Confidence
Lift

A word of caution

The set of extracted rules can be very large!
Do not inspect or display all rules in that case - always use a subset of rules or use the functions head or tail!

Groceries dataset

Let's go back to the Grocery store

Groceries

Dataset from arules package

# Loading the arules package
library(arules)

# Loading the Groceries dataset
data(Groceries)

summary(Groceries)

Summary of Groceries

transactions as itemMatrix in sparse format with
 9835 rows (elements/itemsets/transactions) and
 169 columns (items) and a density of 0.02609146 

most frequent items:
      whole milk other vegetables       rolls/buns             soda           yogurt 
            2513             1903             1809             1715             1372 
         (Other) 
           34055 

element (itemset/transaction) length distribution:
sizes
   1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16   17 
2159 1643 1299 1005  855  645  545  438  350  246  182  117   78   77   55   46   29 
  18   19   20   21   22   23   24   26   27   28   29   32 
  14   14    9   11    4    6    1    1    1    1    3    1 

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.000   2.000   3.000   4.409   6.000  32.000 

includes extended item information - examples:
       labels  level2           level1
1 frankfurter sausage meat and sausage
2     sausage sausage meat and sausage
3  liver loaf sausage meat and sausage

Density of Groceries

# Plotting a sample of 200 transactions
image(sample(Groceries, 200))

image_groceries_200

¹ The density of the item matrix is of 2.6%.

Most and least popular items

Most popular items

itemFrequencyPlot(Groceries,type="relative",
                  topN=10,horiz=TRUE,col='steelblue3')

itemFreqPlot_groceries

Least popular items

par(mar=c(2,10,2,2), mfrow=c(1,1))
barplot(sort(table(unlist(LIST(Groceries))))[1:10],
        horiz = TRUE,las = 1,col='orange')

itemFreqPlot_groceries_leas

Cross tables by index

Contingency tables

# Contingency table
tbl = crossTable(Groceries)
tbl[1:4,1:4]

            frankfurter sausage liver loaf ham
frankfurter         580      99          7  25
sausage              99     924         10  49
liver loaf            7      10         50   3
ham                  25      49          3 256

Sorted contingency table

# Sorted contingency table
tbl = crossTable(Groceries, sort = TRUE)
tbl[1:4,1:4]

                whole milk other vegetables rolls/buns soda
whole milk             2513              736        557  394
other vegetables        736             1903        419  322
rolls/buns              557              419       1809  377
soda                    394              322        377 1715

Cross tables by item names

Contingency tables

# Counts
tbl['whole milk','flour']

[1] 83

# Chi-square test
crossTable(Groceries, measure='chi')['whole milk', 'flour']

[1] 0.003595389

Contingency tables with other metrics

crossTable(Groceries, measure='lift',sort=T)[1:4,1:4]

                 whole milk other vegetables rolls/buns      soda   
whole milk               NA        1.5136341   1.205032 1.571735
other vegetables  1.5136341               NA   1.197047 0.9703476 
rolls/buns        1.2050318        1.1970465         NA 1.1951242 
soda              0.8991124        0.9703476   1.195124        NA

MovieLens dataset

MovieLens: Web-based recommender system that recommends movies for its users to watch.

movielens

Let's watch movies!

Análise de Cesta de Compras em R