Course talk:CPSC522/Gender Classification using Temporal Patterns in Movie Ratings

I thought your article was interesting. One thing you could add is information on the baseline model in the starter.py code (predicting mean). This would give a reference point for how well your model worked. Additionally, it might be helpful to add in a brief one or two sentence overview on LSTM models and why you think they would work well in this context

Thanks. I added some information about what you mentioned in method section.

It would have been better if you had provided some information about LSTM and its formulation. Besides, the reason that makes you think they are a good choice for this problem should be clearly stated. What are "padding and truncating techniques"? If you do not want to explain them explicitly, I think you have to refer to a paper or add a link to help the reader understand these concepts. Stating some of the hyperparameters that you have checked before getting those results can help make your model more reliable for the reader. It is helpful to know the effect of regularization on your model too.

Thanks. I added some information about what you mentioned in method and results sections.

I think the approach is really interesting! Any programming to do with time is always a pain to deal with. It would be cool to see a part on the challenges of working with temporal data - what assumptions do you have to make and which do you have to let go of? A bit late I know, but it came to mind while reading.

Thanks. I added one section about that.