Course talk:CPSC532:StaRAI2020:Prediction on Movie-Lens Dataset

I really like your idea of making the graph. Here I have two feedbacks:

1) It would be great to add more details about your experiment's properties. How large was your dataset and test set? How did you choose the hyperparameters?

2) It could be also nice to have a brief comparison with the previous or similar works with the same idea. Is their result is similar to yours?

We need a better description of the models. Be more explicit about the details. We should know exactly what the models are from reading the description. It should be at a high-level but precise.

What does accuracy mean? (Why that measure?) How does it compare to predicting mean or median?

Thanks for the feedback. I have added the required information.

Thanks for the questions, David. I added the mode-prediction. It ends up putting our models on a very unfortunate light. That's definitely a lesson learned for future times...

Great page! It'd be nice if you can add a small line or two on why you think the multi-hot vector encoding worked better than the one-hot encoding

Thanks for the feedback. I have added the reason for this in page.