machine learning - Recommender system with time window -
what procedure in item-item cf if want calculate probability or predict count of product purchases user in next 7 or ten days based on purchasing history?
for example if user1 buys 2 breads week, probability he/she buy 2 breads after 7 days again.
i thinking of model user-item matrix contain number of purchases , item has feature parameter such week number corresponding time of product purchases, not sure how is.
thanks lot!
Comments
Post a Comment