a generalized recommendation system of Items and ratings, with rating prediction
This project is an inquisitive undertaking on the concepts of Recommendation Systems, Collaborative Filtering, and Item Rating Prediction using collaborative filtering techniques. For this, a sample dataset was acquired to conduct the exploration on various techniques.The Given data set is a user ratings of items. The dataset consists of around 85,000 entries of userID, ItemID, their corresponding rating, and the rating time in timestamps. Using the given data a collaborative filtering model to predict the ratings of given testing data is initiated. The process involves building a User vs Item matrix with ratings as the content, normalizing the matrix, determining the similarities of the entries, and predicting the unknown rating using the known users, items, and similarities. Details on various assumptions made, approaches considered, and the final model used for predicting the ratings of the users are discussed further.