MARST: Multi-Agent Recommender System for e-Tourism Using Reputation Based Collaborative Filtering

Abstract

Abstract. This paper presents a Multi-Agent Recommender system for eTourism (MARST) for recommending tourism services to the users. This system uses Reputation based Collaborative Filtering (RbCF) algorithm that augments reputation to existing Collaborative approach for generating relevant recommendations and to handle cold-start new user problem in tourism domain. The structure of a tourist product is more complex than a book or a movie and hence user profile modeling for these systems is much harder than most of other applications domains like books or movies. Moreover the frequency of activities and rating in tourism domain is also much smaller than in most of the other domains. This increases the complexity in designing and development of Recommender Systems in tourism domain. An attempt has been made in this paper to generate relevant services for a user in tourism domain using reputation based collaborative filtering. Most of the existing Recommender systems focus on one service at a time, whereas the proposed system incorporates three services (hotels, places to visit and restaurants) at a single place to ease the searching of information at one place only. The prototype of MARST has been designed and developed using various JAVA technologies and its performance was evaluated using precision, recall and F 1 metrics.

Analysis and prediction of hotel ratings from crowdsourced data

“…By default, tourism crowdsourcing platforms store MC information regarding multiple service dimensions, for example, comfort, cleanliness, staff. On the one hand, SC profiling remains popular (Bedi, Agarwal, Jindal, & Richa, ; Davoudi & Chatterjee, ) and, in the case of hotels, usually corresponds to the overall rating (Colace, Santo, Greco, Moscato, & Picariello, ; Hu, Lee, Chen, Tarn, & Dang, ; Qian, Feng, Zhao, & Mei, ; Zheng, Burke, & Mobasher, ). On the other hand, the matching between a tourist and a tourism resource may depend on more than one aspect.…”

A Group Recommendation System for Movies based on MAS

“…Multi-agent systems (MAS) have been applied in various domains. When it comes specifically to recommendation systems, some approaches have proposed multi-agent techniques to generate recommendations to both individual users and groups in different domains, like adaptive customization of websites (Morais, et al, 2012), e-commerce (Lee, 2004), games on mobile phones (Skocir, et al, 2012), semantic knowledge extraction (Lopes, et al, 2009), tourism (Bedi, et al, 2014), among others. One thing to notice is that most of those systems can produce recommendations targeted only to individual users.…”

“…The system presented in (Bedi, et al, 2014), MARST, uses a Reputation based Collaborative Filtering (RbCF) algorithm for generating relevant recommendations to a user. Finally, in (Marivate, et al, 2008) the authors present a Multi-Agent approach to the problem of recommending training courses to engineering professionals.…”

Parallel context aware recommender system using GPU and JCuda

“…Hybrid recommender systems use combination of collaborative filtering and content based filtering. Recommender systems are widely used in the different areas such as e-commerce, tourism, news recommendations [4], [5], [6]. There are a lot of e-commerce sites which use recommendation algorithms to make effective recommendation to their users e.g.…”