After a crystal-clear overview of what types of recommender systems (RSs) exists, the simplest one (based on association rule mining) is built from scratch in Python, along with (product) content and (demographics) knowledge-based alternatives. Then some collaborative filtering RSs (both item-based and user-based) using kernel neural networks and support vector machines (using matrix factorization, singular value decompositions and alternative least squares) are developed, along with more advanced hybrid, clustering, classification, deep learning, natural language processing and graph-based RSs. A closing overview of current (as of 2022) emerging techniques as personalized/context-based engines and multi-objective RSs is also included, with the main goal through the whole book being to fully be able to understand and implement different RSs techniques using Python.

https://link.springer.com/shop/apress/titles/en-gb (to appear)

Kulkarni, A.R; Shivananda, A; Kulkarni, A; Krishnan, V.A. Applied Recommender Systems with Python. APress, December 24th, 2022. ISBN 9781484289532.

Pablo Guerrero-Garcia

University of Malaga