• azure
• collaborative filtering
• databricks
• DeepFM
• factorization machine
• lakehouse
• recommender system

Over the past decade, casino gaming has boomed in Italy, especially for online casinos, where virtual casino games currently dominate the market share within the domain. Given a vast array of online casino games available, recommender systems have become increasingly essential for offering personalized game suggestions to users. The project focuses on the development of a recommeder system for potential clients in the betting industry with the goal of providing tailored game recommendations to enhance users experiences. The recommender system was implemented using the Microsoft Azure Cloud and Databricks platforms. Three popular algorithms—ALS, FM, and DeepFM—were evaluated and compared. The experiments used two data split strategies: Global Temporal Spit and Leave One Last Item Split. The results show that DeepFM outperforms the other two models across most evaluation metrics when using the first split strategy, while all three algorithms demonstrate distinct strengths when applying the second strategy.