Redo research to improve the recommendation model without being ready, being a model

[sc0v0ne]

Before improving the model, it will be necessary to improve the movie query, using regex or pandas query, to avoid wrong names and end up generating an exception.

I think it's better to break down this card, as the idea of this card would be to stop having the dataframe already with the groups and separate it with a trained model. Where the user will place the type categories, 3 options. Therefore these 3 choices will be 1,1,1 and the rest 0,0,0. To match the k-means record labels and recommend films based on that. Carry out some tests with the notebooks based on this.

[sc0v0ne]

Use this example: Kaggle Notebook