Publisher's Synopsis
Collaborative filtering (CF) is a popular recommendation approach that has been extensively researched over the last two decades, resulting in a diverse set of algorithms and a large collection of tools to evaluate their performance. This research proposes a new recommendation approach to deal with the problems of grey sheep and data sparsity, with the aim of improving prediction accuracy by inferring new users from existing users in datasets. This transformation creates users with preferences opposite to those of real users, thereby increasing the number of users and solving the two problems mentioned. The performance of this approach has been evaluated using two datasets, MovieLens and FilmTrust. Overall, this book contributes to the development of better recommender systems capable of overcoming the challenges of data overload and improving user experience.
