Recommender Systems and Attributed Networks
Françoise Fogelman‐Soulié, Lanxiang Mei, Zhang Jianyu, Yiming Li, Wen Ge, Li Yinglan, Qiaofei Ye · 2020
Online experience is systematically enhanced through recommender systems (RSs) bringing to users recommendations for movies, products, songs, friends, banners or content on social sites or travels. This chapter introduces RSs and the major algorithms to produce recommendations. It presents social networks and their properties and discusses the particular case of bipartite networks. The chapter also presents new models for recommendation using social networks and discusses the problem of using attributes for enhancing model quality. There exist two main families of techniques to build RSs, depending on the available data: supervised model-based and neighborhood-based models. The latent-space approach has been shown to perform best for rating prediction but neighborhood-based techniques are best for Top-N recommenders with implicit data. Social network analysis allowed us to derive hybrid RSs using attributes, predict a range of optimal N for Top-N recommenders. Filtering low items similarities and representing data in low-dimension spaces are key factors for obtaining performances for RSs.