Collaborative filtering - Regularization with binary labels

[Jesus_CMT]

Hi. I’m watching the lesson of Binary labels: favs, likes and clicks, for Collaborative filtering. Here, we’re given the cost function for cases where the input is numbers:

And then we’re given J for when the input is binary

However, notice the lack of regularization in the second function. Do we not use regularization in this case? Or was it just omitted for simplicity?

[rmwkwok]

Hello @Jesus_CMT,

Nice catch!

Yes, I think the lecture just didn’t add regularization to the cost function for binary predictions. We can certainly add regularization in the same way as your first cost function for numerical predictions.

Cheers,
Raymond