I am trying to model my users' food preferences so that I can recommend restaurants which he/she might be satisfied. The following is some sample data:
uid categories action count
----- ------------------------------------------------- ------- -----
user1 American display 5
user1 American open 2
user1 Irish display 5
user1 Irish open 1
user1 Italian display 6
user1 Local Flavor display 6
user1 Seafood display 6
user2 Barbeque display 6
user2 Barbeque open 1
user2 Coffee & Tea display 2
user2 Local Flavor display 6
user2 Pizza,Italian display 2
user2 Sandwiches,Salad,Soup display 4
user2 Sandwiches,Salad,Soup open 2
user2 Seafood display 1
user2 Sushi Bars,Japanese display 2
As you can see, we can calculate the click through rate (CTR) by the counts of action 'display' and 'open', and I want to maximize the CTR for the recommendations.
Since this is a secondary function on a mobile app, the conventional recommendation techniques like SVD or ALS which needs heavy computation might not be appropriate. Is there any other approach so that I can model and predict the users' preference?