I am trying to build a cross selling model where I have customers and some of their attributes along with the products that they have bought. SO basically I will be having multiple response variables depending on the product he bought and the quantity. I need to predict amongst all the products, what is the probability of each product being bought next(in the future). Which algorithm (random forest, naive bayes etc.. )can model this and handle multiple response variables in separate columns?
asked Oct 25, 2017 at 17:22
-
1Please ask at Stats or Data Science in more specific example.– CallMeLaNNCommented Oct 25, 2017 at 20:16
-
Sure. Sorry about that. I wasn't aware– Digvijay SawantCommented Oct 26, 2017 at 18:20
Add a comment
|
1 Answer
You can very well use Neural Networks or Random Forests for your problem. I have already modelled similar kind of problems using these two machine learning models. Below is my git link for the sample code to implement Neural Networks and Random Forests. https://github.com/naveenkambham/826Project If you think it is helpful mark/vote
answered Oct 25, 2017 at 18:38
-
Thank you so much. KNN is also another option. It has the ability to take in multiple columns as response. I think NN and random forest might work too. Commented Oct 26, 2017 at 2:44