Effective recommendation systems rely on capturing user preferences, often requiring incorporating numerous features such as universally unique identifiers (UUIDs) of entities. However, the exceptionally high cardinality of UUIDs poses a significant challenge in terms of model degradation and increased model size due to sparsity. This paper presents two innovative techniques to address the challenge of high cardinality in recommendation systems. Specifically, we propose a bag-of-words approach, combined with layer sharing, to substantially decrease the model size while improving performance. Our techniques were evaluated through offline and online experiments on Uber use cases, resulting in promising results demonstrating our approach's effectiveness in optimizing recommendation systems and enhancing their overall performance.

Recommendation systems can be defined as software applications that draw out and learn from data such as preferences, their actions (clicks, for example), browsing history, and generated recommendations, which are products that the system determines are appealing to the user in the immediate future. In this tutorial, we will learn to build a recommendation system with Scala and Apache Spark. This article is an excerpt taken from Modern Scala Projects written Ilango Gurusamy. In the preceding diagram, can be thought of as a recommendation ecosystem, where the recommendation system is at the heart of it. Implementation is documented in the following subsections.