Download the file Movie Ratings (.csv) at the bottom of the page of chapter 2. You will note that a lot of the values a values are NaN (movies rated on a scale of 1-5; a blank means that person didn’t see that movie).
Build a recommendation system using k-nearest neighbor approach (similar to the one presented in the book). using 2 cases:
- ( case ) You should use cosine similarity for a distance measure instead.(in the book Pearson correlation is used for a distance)
- ( case) Fill in the missing values with mode of the ratings (each column has a different mode ) . Use euclidean distance for a measure
You should submit / upload one IPython Notebook file ( .ipynb )