Alejandro Bellogin

Alejandro Bellogín is an Associate Professor at Universidad Autónoma de Madrid. Previously, he held a post-doctoral research grant with the Centrum Wiskunde and Informatica in Amsterdam, The Netherlands. He has supervised around 10 Master Theses, 1 PhD thesis, and more than 20 Bachelor Theses on recommender systems. His publication history includes around 80 publications about different aspects of recommender systems.

projects by Alejandro Bellogin

Similarities and Recommender Systems

4 weeks · 4-6 hours per week average · BEGINNER

In this series of liveProjects, you’ll complete all the steps to build a recommendation system for a movie streaming website. Your system will need to learn the tastes of its users and predict which movies they might want to watch. You’ll work with movie data points like names, genres, and directors, and mapped interactions such as ratings and tags.

Learning Similarities

1 week · 4-6 hours per week · BEGINNER

In this liveProject, you’ll determine what similarities there are between certain movies, and then build a recommendation system based on the similarities you’ve identified. Although you’ll be working with movie data for a streaming website, identifying similarities can help enhance recommendation engines for any item or product. You’ll develop a general technique for spotting similarities, then apply your learned similarities to build a recommendation engine. You’ll then enrich your model with collaborative and content-based data, and evaluate and compare your model.

Dimensionality Reduction

1 week · 4-6 hours per week · BEGINNER

In this liveProject, you’ll build a movie recommendation system that reduces the dimensionality of the spaces involved to enhance its performance so that latent representations are identified, which aim to capture users and items interactions with the system on the same latent vector space. You’ll work to compute clusters on the movies and clients of your site, and utilize latent representations of users and items with a reduced dimensionality to compute similarities. Once you’ve established this data, you’ll build and test your recommendation engine.

Content-Based Similarities

1 week · 4-6 hours per week · BEGINNER

In this liveProject, you’ll build a movie recommendation system based on the content and metadata of movies in your system. This system is intended to maximize the satisfaction of your movie-watching users. You’ll start with an analysis to determine the content of your movies, then use that data to implement content-based similarities for both products and users. You’ll build and evaluate your recommender system based on these connections, till it’s the best it can be!

Collaborative Similarities

1 week · 4-6 hours per week · BEGINNER

In this liveProject, you’ll build a movie recommendation system based on movies that your users have previously interacted with. This system is designed to boost engagement and keep your users on your site. You’ll develop a competitive array of similarity functions, and create your own recommender system based on these similarities. Finally, you’ll evaluate your system’s effectiveness and tune its parameters.