implementation of RSTE

[hdi-amiri]

I have one question about implementation of RSTE algorithm, why there is no difference between update rule of this model and PMF? how we have exploited the social data in updating parameters?

[Coder-Yu]

In the function named predict, you can see the social data is considered. Then the prediction is returned for training. Therefore, the error between the known rating and the prediction contains the social information. That makes the implementation of RSTE different from that of PMF.