Personalized Hybrid Book Recommender

Personalized Hybrid Book Recommender

Hossein Arabi (Faculty of Computer Science and Information Technology, University of Malaya, Kuala Lumpur, Malaysia) and Vimala Balakrishnan (Faculty of Computer Science and Information Technology, University of Malaya, Kuala Lumpur, Malaysia)
Copyright: © 2019 |Pages: 28
DOI: 10.4018/IJISSS.2019070105

Abstract

Personalized Recommendation Systems (RS) provide end users with suggestions about items that are likely to be of their interest based on users' details such as demographics, location, time, and emotion. In this article, a Personalized Hybrid Book Recommender (PHyBR) is presented, which integrates personality traits with users' demographic data and geographical location to improve the quality of recommendations. The Ten Item Personality Inventory (TIPI) was used to determine users' personality traits. PHyBR was evaluated using two metrics, that are, Standardized Root Mean Square Residual (SRMR) and Root Mean Square Error of Approximation (RMSEA). Both metrics revealed PHyBR outperforms the baseline models (without considering personality traits and geographical location factor) in terms of the recommendation accuracies. This study shows that users who are in the same geographical contexts intend to have similar preferences. Therefore, users' personality details along with their geographical locations can be used to provide improved personalized recommendations.
Article Preview
Top

1. Introduction

With the growing penetration of the Internet and e-commerce, personalized recommendations that identify appropriate products or services for customers to reduce their information load and cost of searching become increasingly critical. Many online vendors, including Amazon and Netflix, have implemented Recommendation Systems (RSs) to assist their consumers (Chen et al., 2010). However, central to the development of an effective RS is identifying customer preferences, which can be analyzed by various criteria such as customer browsing behavior, personality or purchasing history (Adomavicius & Tuzhilin, 2005).

RSs normally acquire data about user’s activities and build user models to filter the preferences expressed either explicitly (observing rating or wish list) or implicitly (monitoring user’s behavior such as websites visited, songs heard and books read) (Núñez-Valdéz et al., 2012). The recommender technology is superior to other information filtering applications because of its ability to provide personalized and meaningful information recommendations. For example, while standard search engines are very likely to generate the same results to different users entering identical search queries, RSs are able to generate personalized results which are more relevant to the user as they take each user’s personal interests into account (Gavalas et al., 2014).

As users of RSs may have different needs in various situations and contexts, it is becoming increasingly important to consider contextual data when filtering information (Hawalah & Fasli, 2014). This resulted in the birth of personalized recommendations, focusing on various user contexts such as time of access (Wang & Shao, 2004), location of access (Braunhofer et al., 2014; Huang, 2016; Liu et al., 2013) and emotion/mood (Shan et al., 2009). For instance, studies have shown significant correlations between personality and people’s tastes and interests. For example, Cantador et al. (2013) revealed similar personality traits between users who like science-fiction and comic books. Similarly, those who like self-help books show strong personality resemblance to those who like mystery books. A more recent study used users’ Facebook profiles to determine their personality traits in order to help improve book recommendations (Bhosale et al., 2017). Pera et al. (2011) on the other hand, developed a book recommender based on social interactions and personal interests to suggest books to users. Zhang et al. (2013) combined users’ virtual ratings and real ratings to improve recommendation accuracy using the Amazon book dataset. The relationship between users' implicit tastes (i.e. what the user likes) and products' inherent properties (i.e. what the book is about) was examined to improve book recommendations in McAuley and Leskovec (2013). Finally, Chen (2013) developed a mobile location-aware book RS that uses map navigation to recommend books to learners within a real-library environment.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 12: 4 Issues (2020): 2 Released, 2 Forthcoming
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing