We use A Priori Algorithm as association rule to build the recommendation system. Apriori is a probability based rule. Starts from the market basket transaction, Apriori finds frequent itemsets through a method similar to tree construction
Use minimum support of 0.6 and minimum confidence of 0.9, we get our training results:
|
Then in our app, if you are interested in Mac Chicken, click into the page, will see that people who order Mac Chicken also like Mac Fish.
|