Recommendation System for Movie Ratings (2 problems need to solve)
Budget: ₹600 – ₹3,000 INR
Implementation
Implement the Item-based collaborative filtering algorithm (please do not use any off the-shelf libraries for collaborative filtering, the idea is to implement the algorithm yourself)
• Step 1: Use the ratings specified in ratings.-train.txt file to learn similarity between movies. You can use the correlation coefficient (as given in the slides) or cosine similarity to compute the similarity weights.
• Step 2: For each row in ratings-test.txt, predict the rating based on a neighborhood of similar movies. Refer to the slides on collaborative filtering for the equation used to compute the predicted rating. You can select any appropriate neighborhood size when making the prediction.
Implement the Item-based collaborative filtering algorithm (please do not use any off the-shelf libraries for collaborative filtering, the idea is to implement the algorithm yourself)
• Step 1: Use the ratings specified in ratings.-train.txt file to learn similarity between movies. You can use the correlation coefficient (as given in the slides) or cosine similarity to compute the similarity weights.
• Step 2: For each row in ratings-test.txt, predict the rating based on a neighborhood of similar movies. Refer to the slides on collaborative filtering for the equation used to compute the predicted rating. You can select any appropriate neighborhood size when making the prediction.