I have this weird kind of requirement that i need for my azure ml model, not sure how to make it, so its that i will provide the model a bunch of data and in that data each row signifies a single interaction of a user with my app. Now the same user may have multiple interactions too so more than one rows is possible for same user. Also a user is uniquely determined by the customer_id column. Now the other columns describe the details of that user interaction, so for eg lets say my app is a shopping app, so the other columns could be like - price of the item, rating of the item, discount etc.

Also i have one special column called the level column which tells me how much further did user go with that item, like did he just click on it or added it to cart or did he buy it and so obviously buying in the level column is a stronger indicator of preference than just clicking. So i want the model to go through all these user interactions data and prepare a new table which should have this - against every customer id, a list of features/columns that are most important to him/her based on his/her interactions.

Finally I want to create something like a rest api web service from this model, where i just provide it the customer id and it will give me the list of features/columns that will be important for this user. Need guidance regarding how to implement this using azure ML studio?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.