how do you make that calculation to determine the best (top matching) rule for market basket

sky

You find top matching rules in term of highest support / highest confidence, and so you determine the associated Conclusions and thus the recommendations.

 

My attached image has the support and confidence %. But how do I find the top matching rules? What are the customer recommendations based on the data? also, can we make recommendations based on "lift" as well? THank you

Maerkli

Hallo Sky,

 

If you have any time, please, read Chapter 6 of Predictive Analytics and Data Mining
Concepts and Practice with RapidMiner by Vijay Kotu & Bala Deshpande, PhD.

You can find easily a PDF version.  Association Analysis is explained in detail.

Maerkli.