08_collab - Question

Just a quick question- in chapter 8, with the movie lens data, shouldn’t the dot product of latent factors result in movie ratings? I am not sure if I am missing something here but the data I am getting is oddly different to what was initially fed as input to the model.

Original ratings:

Solved - took a while to figure out but finally got there. I had to convert the dot product back into the ratings using sigmoid_range.

Here is the screenshot.

1 Like