ABSTRACT
The increase in online activities is directly proportional to the ubiquitous presence of recommender systems. User experience and ease of use on a system is a key factor for an efficient online system, and so recommender systems are quite inevitable. Recommender systems are one of the very useful and widely used techniques in machine learning (ML)/artificial intelligence (AI) and have been a pivotal technique in the application of ML/AI technologies in business processes. ML/AI algorithms are adopted to achieve a highly efficient recommender system as there exist a large sample of data. This chapter covers the machine learning algorithms that are associated with recommender systems. It also highlights the hybridization of these algorithms and how robust solutions are achieved from it. It covers factorization-based methods that are very useful in reducing the resultant huge matrices on the space of a recommender system. The chapter also discusses the predictive functionality of memory-based model and model-based methods and their applications. Consequently, trainers such as autoencoders and deep autoencoders are discussed as well as its application with case studies. Deep and reinforcement learning algorithms and statistical methods applicable to recommender systems are also highlighted.
