I have customer purchase history and some independent variables associated with each product that customer bought. I am trying to predict the next best predict for customers.
T Customer Input1 Input2 ...... Input10 PrdouctBought 1 cust1 2 cust1 3 cust1 4 cust1 1 cust2 2 cust2 3 cust2 4 cust2
I have restricted each customer purchase history to 4 so that we can fix batch_size=4 and we get away with the problem of model taking in account of last product of previous customer. What I am not able to do is how to incorporate these input variables to make RNN predict the next best product for that customer.