Application of Machine-Learning Techniques in the Development of Neighbourhood-Based Robust Recommender Systems

Document Type

Book Chapter

Publication Title

Recommender Systems: A Multi-Disciplinary Approach

Abstract

Recommender systems (RSs) are a type of information processing tool that uses a lot of different types of information filtering processes to figure out what a customer wants and to give them relevant information. There are numerous statistical machine-learning techniques that can be utilised to better comprehend the principles and challenges of RSs. This chapter aims to look into these machine-learning techniques and the mechanisms of their involvement in this context. RS techniques are often divided into three categories: collaborative filtering (CF), content-based filtering (CBF), and hybrid. We will mainly talk about the techniques used in the CF method that allow users to discover new content that is different from what they have seen before. This chapter evaluates and discusses the utility of traditional network clustering techniques such as Louvain, Infomap, and label propagation algorithms for the development of neighbourhood-based robust RSs. We also look into and incorporate a nodality-based network clustering method to make another neighbourhood-based robust RS. This chapter mainly discusses a strategy for building robust RSs that integrates a network clustering approach with neighbourhood-based RSs, especially the adsorption algorithm. Extensive experimental assessments on real world datasets demonstrate the utility of the integrated neighbourhood-based RSs.

First Page

203

Last Page

234

DOI

10.1201/9781003319122-13

Publication Date

1-1-2023

This document is currently not available here.

Share

COinS