Anallyz Shopping Cart

 

 

Self-use Tool

 

Sample Data

 

We will see an interesting and most commonly used application of "Association Rule Mining". Please download the Sample Data, which has close to 10000 Transactions from a Retail Outlet on Grocery items. 

 

Please use the Self-service Tool, and use "Association Rules" option. It would give us three interesting insights. 

 

  • What are the "Frequent Item Sets"?
  • What are the Strongest Rules? 
  • What are most associated items for a given Item of choice? 

 

Please load the Sample Data in the tool and see the first result on "Frequent Items". The 10 most Frequent Items are shown in the below picture. Similarly one can input other numbers to view different number of Items. Whole Milk, Other Vegetables, Soda etc are the most frequent items based on shear number of their occurrence in the Transaction Data. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Please use "View Rules" option, with "Minimum Support", "Minimum confidence" and "Number of Rules" to list the strongest rules. Usually the default values should work, but decrease Support and/or Confidence to view rules if required. The following picture on left shows five strongest rules, with their Support, Confidence and Lift. The LHS and RHS should be interpreted as if-then. As an example, the top rule in the picture should be read as if "Rice-Sugar" then "Whole Milk". 

 

Sometimes, we may be interested in knowing associated items for a particular item. As an example, we may be interested in knowing the most frequently bought items with "whole milk". Please use "Analyze an Item" option to see list of associated items. There are two ways to analyze an item, "Before" and "After". The following picture on right shows the "After" rules for "whole milk" (case sensitive). The "before" rule would throw the same results as on the left side. You may want to check the result yourself.