An empirical study of natural noise management in group recommendation systems
Introduction
A recommender system (RS) focuses on providing personalised access to items (or information in general) in an overloaded search space. With this purpose in mind, RSs explore users' interests to find out which items are the most suitable. RSs have been successfully applied to support users trying to overcome the information overload problem in several domains [15], such as e-commerce [7], financial investment [20], e-learning [16], [28], e-government [14], and e-tourism [22], among others.
RSs employ different approaches depending on the information they rely on, but the most widespread ones are content-based and collaborative filtering approaches [1]. A content-based recommender system (CBRS) relies on users' preferences and information about items, such as features or a textual description, to be able to make recommendations [10]. On the other hand, collaborative filtering recommender systems (CFRSs) rely on users' preferences to be able to make recommendations [21]. Therefore, CFRSs are able to generate effective recommendations using only users' preferences (rating values) [25] hence the quality of the ratings influences the quality of the recommendations.
The ratings can be gathered implicitly [6] or explicitly, this article focuses on the latter case. Prior research has explored how noisy preferences intentionally inserted by users affect RSs [8], [13], so-called malicious noise. However, the noisy ratings introduced unintentionally by users, so-called natural noise, has recently attracted the attention of researchers. Several proposals investigate the detection and correction of such natural noise. Some proposals perform these tasks by exploiting the items' attributes [24], either by taking advantage of user's interaction [3], or by using knowledge extracted from the ratings themselves [29]. The main benefit of these proposals is their positive impact on the recommendations.
So far, natural noise has been studied only in RSs for individuals. However, Group Recommender Systems (GRSs) [11] play an important role in many social activities that require recommendations to be delivered to a group of users, such as watching TV with family, sightseeing with others, or going to the cinema with friends.
GRS approaches extend individual RS for recommending to groups by aggregating user individual information [18]. Therefore, GRSs use explicit ratings, and natural noise is also present biasing the group recommendation. Consequently, its management may play an important role in the quality of the group recommendations. This paper aims at researching the natural noise management (NNM) in GRS to study its influence in group recommendation.
In group recommendation, unlike traditional recommendation scenarios, the ratings dataset has different levels of information from the group viewpoint. Therefore, the direct application of NNM methodologies is not adequate in this context [29]. This fact makes necessary to develop new methodologies that enable the application of NNM across the different levels of information in GRSs. This is the objective of the current paper, which will be reached by the study of different hypotheses.
Taking into account that group members' ratings are key data in GRSs for computing the recommendations provided to the group by their aggregation [11], their quality influences the recommendation accuracy. It leads to the formulation of the first hypothesis:
- •
H1. NNM using only the group ratings would improve the group recommendation.
On the other hand, NNM for individual RSs has shown clear improvements [29]. Due to the fact that, GRSs approaches are supported by individual RSs [11], a new hypothesis is raised:
- •
H2. NNM in the entire ratings database, disregarding the groups, would improve the group recommendation.
Eventually, if both levels (group ratings, all users ratings) are considered, some ratings tagged as noisy at the group level could be not noisy at the global level, and vice versa. Therefore, the NNM at both levels could lead to a better recommendation accuracy. Hence, H3 is formulated as follows:
- •
H3. managing natural noise in the entire ratings database and, after that, adding a second step that manages natural noise in the group ratings, would improve the results as compared to a single step of NNM.
Different approaches which apply these NNM processes to GRSs are presented and a case study to verify the hypotheses is performed.
The remaining of the paper is organised as follows. Section 2 provides a background on CFRSs, GRSs, and natural noise. Section 3 presents four methods of natural noise management for GRSs. Section 4 develops the case study and analyses the results of the proposals, leading to the acceptance of H2 and H3. Finally, Section 5 points out the conclusions and discusses future research.
Section snippets
Background
This section briefly reviews the basics on CFRSs and GRSs. Eventually, some research work on NNM is discussed.
Natural noise management in group recommendation
The NNM is particularly interesting in GRSs, because it is not clear whether the natural noise of the members' ratings also bias the group recommendation. Therefore, it is important to verify that the NNM also plays an important role in GRS accuracy. However, NNM in individual RSs cannot be directly applied to GRSs, because of the proper features of GRSs. Therefore, different alternatives for NNM in group recommendation are introduced in this section.
To propose a NNM approach in GRS, in the
Case study
To measure the effect of previous NNM approaches in GRSs performance, an experimental procedure is used to evaluate them. This section presents the experimental protocol used in the experiments, and the results are then presented and discussed to verify the hypotheses introduced in Section 1.
Concluding remarks and future works
This paper presents four approaches to manage and correct the natural noise in group recommender systems. The results show that NNM over the group ratings provides slight improvements to the group recommendation performance. On the other hand, when the NNM is applied to the entire dataset, it clearly increases the performance of the group recommender systems. The best results were obtained by the NNM-H approach, which performs a cascade hybridisation of the global and local approaches, i.e., it
Acknowledgments
This research work was partially supported by the Research Project TIN-2015-66524-P, the Spanish FPU fellowship (FPU13/01151), and also the Eureka SD Project (agreement number 2013-2591).
Jorge Castro received the BSc. and MSc. degrees in Computer Science from the University of Jaén, Spain. In 2011 he started in research in the same university, in which he collaborated in several research projects. He is currently a PhD student in the Department of Computer Science and Artificial Intelligence in the University of Granada, Spain. During his PhD, he is also developing a software library focused on experimenting with recommendation algorithms. His research interests are recommender
References (30)
- et al.
Recommender systems based on quantitative implicit customer feedback
Decis. Support. Syst.
(2014) - et al.
Recommender system application developments: a survey
Decis. Support. Syst.
(2015) - et al.
Personalized finance advisory through case-based recommender systems and diversification strategies
Decis. Support. Syst.
(2015) - et al.
A mobile 3d-GIS hybrid recommender system for tourism
Inform. Sci.
(2012) - et al.
Toward the next generation of recommender systems: a survey of the state-of-the-art and possible extensions
IEEE Trans. on Knowl. Data Eng.,
(2005) - et al.
I like It.. I like It Not: Evaluating User Ratings Noise in Recommender Systems
- et al.
Rate It Again: Increasing Recommendation Accuracy by User Re-Rating
- et al.
Data Mining Methods for Recommender Systems
- et al.
Group recommendation: semantics and efficiency
Proceedings of the VLDB Endowment
(2009) - et al.
E-commerce recommendation applications
Data Mining and Knowledge Discovery,
(2001)
Robust Collaborative Recommendation
Weighting of features in content-based filtering with entropy and dependence measures
International Journal of Computational Intelligence Systems
Semantics-Aware Content-Based Recommender Systems
Comparison of group recommendation algorithms
Multimed. Tools. Appl.
E-learning recommender system for a group of learners based on the unified learner profile approach
Expert. Syst.
Cited by (44)
Sampling and noise filtering methods for recommender systems: A literature review
2023, Engineering Applications of Artificial IntelligenceA similarity-enhanced hybrid group recommendation approach in cloud manufacturing systems
2023, Computers and Industrial EngineeringA fuzzy content-based group recommender system with dynamic selection of the aggregation functions
2022, International Journal of Approximate ReasoningSEDGN: Sequence enhanced denoising graph neural network for session-based recommendation
2022, Expert Systems with ApplicationsLess is more: improving neural-based collaborative filtering by using landmark modeling
2022, Information Sciences
Jorge Castro received the BSc. and MSc. degrees in Computer Science from the University of Jaén, Spain. In 2011 he started in research in the same university, in which he collaborated in several research projects. He is currently a PhD student in the Department of Computer Science and Artificial Intelligence in the University of Granada, Spain. During his PhD, he is also developing a software library focused on experimenting with recommendation algorithms. His research interests are recommender systems, group recommendation and context-aware recommendation.
Raciel Yera received the BSc. in Computer Science from the Computer Science University, Cuba in 2010, his MSc degree in 2012 from the University of Ciego de Ávila, Cuba, and his PhD degree in 2015 from Las Villas Central University, Cuba. At present, he is a professor of Computer Science at University of Ciego de Ávila. He obtained three Annual Provincial Awards of the Cuban Academy of Sciences in 2013 for his research results related to recommender systems. His current interests are focused on the improvement of recommender systems performance through the application of computational intelligence techniques.
Luis Martínez (M’10) received the M.Sc. and Ph.D. degrees in Computer Sciences, both from the University of Granada, Spain, in 1993 and 1999, respectively. Currently, he is Full Professor of Computer Science Department and Head of ICT Research Centre at the University of Jaén. His current research interests are linguistic preference modelling, decision making, fuzzy logic based systems, computer aided learning, sensory evaluation, recommender systems and electronic commerce. He co-edited nine journal special issues on fuzzy preference modelling, soft computing, linguistic decision making and fuzzy sets theory and published more than 75 papers in journals indexed by the SCI as well as 30 book chapters and more than 120 contributions in International Conferences related to his areas. In 2015, he authored a monograph The 2-tuple Linguistic Model: Computing with Words in Decision Making (Springer).
Dr. Martínez is a member of the European Society for Fuzzy Logic and Technology and IEEE, Co-Editor in Chief of the International Journal of Computational Intelligence Systems and an Associate Editor of the journals IEEE Transactions on Fuzzy Systems, Information Fusion, the International Journal of Fuzzy Systems, Journal of Intelligent & Fuzzy Systems, the Scientific World Journal, Journal of Fuzzy Mathematics and serves as member of the journal Editorial Board of the Journal of Universal Computer Sciences. He received twice the IEEE Transactions on Fuzzy Systems Outstanding Paper Award 2008 and 2012 (bestowed in 2011 and 2015 respectively). He is also Visiting Professor in University of Technology Sydney, Guest Professor in the Southwest Jiaotong University and honourable professor in Xihua University both in Chengdu (China).