Julia 1.0 Programming Complete Reference Guide

Learning about the model-based approach

This innovative approach to recommender systems was named model-based, and it made extensive use of matrix factorization techniques. In this approach, models are developed using different machine learning algorithms to predict a user's ratings. In a way, the model-based approach can be seen as a complementary technique to improve memory-based recommendations. They address the matrix sparsity problem by guessing how much a user will like a new item. Machine learning algorithms are used to train on the existing vector of ratings of a specific user, and then build a model that can predict the user's score for an item that the user hasn't tried yet. Popular model-based techniques are Bayesian Networks, ...