Collaborative filtering toolkit

User 80 | 3/2/2014, 7:29:58 AM

Hi Danny,

I have a couple of questions that I want ask: 1- After I train and build the linear model, could I use it for predictions later? i.e Could I save the model and load it later to make predictions? 2- If the answer of my first question was yes, Could I update my saved trained model incrementally at anytime later?

Thanks for your time and efforts, Hossam

Comments

User 6 | 3/2/2014, 3:31:36 PM

Hi Hossam, 1. In a couple of days we are releasing our Python beta inteface to GraphLab, where you can easily store the model and compute prediction later. See some examples here: http://graphlab.com/learn/notebooks/recsysexplicitrating.html

  1. We do not support incremental models at this point - you will need to compute the model again with the new data.

Best,