Novel automatic group identification approaches for group recommendation

Reproducibility is a challenging aspect that considerably affects the quality of most scientific papers. To deal with this, many open frameworks allow to build, test, and benchmark recommender systems for single users. Group recommender systems involve additional tasks w.r.t. those for single users, such as the identification of the groups, or their modeling. While this clearly amplifies the possible reproducibility issues, to date, no framework to benchmark group recommender systems exists. In this work, we enable reproducibility in group recommender systems by extending the LibRec library, which stands out as one of the richest, with more than 70 different recommender algorithms, good performance and several evaluation metrics. Specifically, we include several approaches for all the stages of group recommender systems: group formation, group modeling strategies, and evaluation. To validate our framework, we consider a use-case that compares several group building, recommendation, and group modeling approaches.

Recommender systems are increasingly used in various domains like movies, travel, music, etc. The rise in social activities has increased the usage of recommender systems in general and group recommender systems in particular. A group recommender system is a system that recommends items to a group of users collectively, given their preferences. In addition to the user preferences, using social and behavioural aspects of group members to generate group recommendations will increase the quality of the content recommended in heterogeneous groups. Group recommender systems also address the cold start problem that arises in an individual recommendation system. This paper presents a survey on the state-of-the-art in group recommender systems concerning various domains. We discussed existing systems with respect to their aggregation and user preference models. This organisation is very useful to understand the intricacies with respect to each domain.

Group Recommendation (GR) is the task of suggesting relevant items/events for a group of users in online systems, whose major challenge is to aggregate the preferences of group members to infer the decision of a group. Prior group recommendation methods applied predefined static strategies for preference aggregation. However, these static strategies are insufficient to model the complicated decision making process of a group, especially for occasional groups which are formed adhoc. Compared to conventional individual recommendation task, GR is rather dynamic and each group member may contribute differently to the final group decision. Recent works argue that group members should have non-uniform weights in forming the decision of a group, and try to utilize a standard attention mechanism to aggregate the preferences of group members, but they do not model the interaction behavior among group members, and the decision making process is largely unexplored.In this work, we study GR in a more general scenario, that is Occasional Group Recommendation (OGR), and focus on solving the preference aggregation problem and the data sparsity issue of group-item interactions. Instead of exploring new heuristic or vanilla attention-based mechanism, we propose a new social self-attention based aggregation strategy by directly modeling the interactions among group members, namely Group Self-Attention (GroupSA). In GroupSA, we treat the group decision making process as multiple voting processes, and develop a stacked social self-attention network to simulate how a group consensus is reached. To overcome the data sparsity issue, we resort to the relatively abundant user-item and user-user interaction data, and enhance the representation of users by two types of aggregation methods. In the training process, we further propose a joint training method to learn the user/item embeddings in the group-item recommendation task and the user-item recommendation task simultaneously. Finally, we conduct extensive experiments on two real-world datasets. The experimental results demonstrate the superiority of our proposed GroupSA method compared to several state-of-the-art methods in terms of HR and NDCG.

A group recommender system (GRS) is a system that collectively recommends items to a group of users based on their preferences. The GRS and the individual RS challenge lies in a very small and incompleteness of user-item ratings. Such incompleteness resulted in the data sparsity problem. The issues of data sparsity in a group negatively affect the quality of recommendations to the group. It occurs due to the inefficient formation of groups, which usually involves individuals with sparse data in their user profiles. Most of the current studies focus on this issue after the formation of groups. However, this study focused before the group formation, based on the intuition that it will be more efficient if the data sparsity at the individual level is addressed before the group formation process takes place. Therefore, applying the approach through Linked Open Data (LOD) technology is proposed to ensure that the data sparsity issues can be overcome before the group formation process is implemented. We proposed a GRS-LOD model. The experimental evaluations relating to the prediction accuracy and recommendation relevancy of the proposed model were implemented on three aspects: comparison with the basic approach or baselines; comparison with the current approaches, and comparison in terms of group size and aggregation strategies. The aggregation strategies used were the Average (AV), Most Pleasure (MP), Average without Misery (AVM), and Least Misery (LM). The metrics for prediction accuracy were based on the RMSE and MAE, whereas for relevancy, precision, recall, and F1-score were considered. The results show that the prediction accuracy and relevancy of the developed model’s recommendations is better than the baseline study by adapting the Average (AV) strategy with the individual profile aggregation approach. Meanwhile, for the evaluation in terms of group size, the results show larger group size exhibits better prediction accuracy for the four used aggregation strategies. On the other hand, in terms of recommendation relevancy, the result shows that relevancy decreases with the increase in group size for the MP, AV and AVM strategies.

Investigating and counteracting popularity bias in group recommendations

Current researches have demonstrated that the significance of Multi-Criteria Decision-Making (MCDM) methods in Group Recommender Systems (GRSs) has yet to be thoroughly discovered. Thus, we have proposed a Multi-criteria GRS (MCGRS) to provide recommendations for group of users based on multi-criteria optimization. The idea behind our approach is that, each member in a group have different opinions about each criterion and he/she would try to make the best use of multi-criteria to fulfill his/her own preference in decision-making process. Therefore, we have employed Analytical Hierarchy Process (AHP) to learn the priority of each criterion to maximize the utility for each criterion. Then, MCGRS generate the most appropriate recommendation for the group. Experiments are performed on Yahoo! Movies dataset and the results of comparative analysis of proposed MCGRS with baseline GRSs techniques clearly demonstrate the supremacy of our proposed model.

Hybrid POI group recommender system based on group type in LBSN

Point of Interest (POI) recommender systems help provide their users with a location or place that they might be interested in visiting. When combined with Location-Based Social Networks (LBSNs), POI recommender systems can be restructured to recommend POI for groups of users and not just individuals. The research focused on Group Recommender Systems (GRSs) and specifically, POI GRSs are scarce when compared to recommender systems for individuals. There are two main techniques that are used for POI GRSs, the Group Profiling methods and the Users Score Aggregation methods. Both methods have their drawbacks, as the Group Profiling methods do not recommend well for new groups and the Users Score Aggregation methods generally do not perform as well as the Group Profiling methods for established groups. In this paper, we propose new result aggregation methods that use both the Group Profiling method and the Users Score Aggregation method's results to provide the best POI group recommendations without the aforementioned drawbacks.

An empirical study of natural noise management in group recommendation systems

Lots of activities, like watching a movie or going to the restaurant, are intrinsically group-based. To recommend such activities to groups, traditional single-user recommendation techniques cannot be adopted, as a consequence, over the years, a number of group recommender systems have been developed. Recommending to groups items to be enjoyed together poses many ethical challenges, in fact, a system whose unique objective is to achieve the best recommendation accuracy possible, might learn to disadvantage submissive users in favor of more aggressive ones. In this work we investigate the ethical challenges of context-aware group recommendations, in the more general case of ephemeral groups (i.e., groups where the members might be together for the first time), using a method that can recommend also items that are new in the system. We show the goodness of our method on two real-world datasets. The first one is a very large dataset containing the personal and group choices regarding TV programs of 7,921 users w.r.t. sixteen contexts of viewing. The second one, which has been collected specifically for this work and that is made publicly available as one of the contributions of this article, gathers the musical preferences (both individual and in groups) of 280 real users w.r.t. two contexts of listening. We compare the results of our approach with seven other group recommender systems specifically developed to be fair. We evaluate the goodness of our recommendations using recall, while their fairness is assessed using two measures found in the literature, namely, score disparity and recommendation disparity. Our extensive experiments show that our method always manages to obtain the highest recall while delivering ethical guarantees in line with the other fair group recommender systems tested.Keywordsgroup recommender systemscontext-aware recommender systemscomputer ethicsfairness

A fuzzy approach for natural noise management in group recommender systems

Group recommender systems are based on aggregation heuristics that help to determine a recommendation for a group. These heuristics aggregate the preferences of individual users in order to reflect the preferences of the whole group. There exist a couple of different aggregation heuristics (e.g., most pleasure, least misery, and average voting) that are applied in group recommendation scenarios. However, to some extent it is still unclear which heuristics should be applied in which context. In this paper, we analyze the impact of the item domain (low involvement vs. high involvement) on the appropriateness of aggregation heuristics (we use restaurants as an example of low-involvement items and shared apartments as an example of high-involvement ones). The results of our study show that aggregation heuristics in group recommendation should be tailored to the underlying item domain.

<p>This research article presents a study about the background in <em>Group Recommender Systems </em>and how <em>social factors </em>are directly related to these applications. Some important group recommender systems in academia are described to exemplify their contribution in different domains. Besides, a framework that is intended to improve group recommender systems is proposed. The main idea of the framework is to enhance social cognition to help the group members agree and make a decision. Its structure includes a process where an influential group is detected among the target groups of people to recommend to. Social influence detection uses the knowledge behind online social connections and interactions. Trying to understand human behavior and ties among groups in a social network and how to use this to improve group recommender systems is considered the main challenge for future research. Combining this with the kind of item recommendation which involves a temporal sequence of ordered elements will present a novel and original path in Group Recommender Systems design.</p>

An improved clustering scheme based on user group is proposed. Every two users are grouped among N-users in the allowed system according to their link gain from large to small. Each user group is numbered sequentially. Two user clusters are obtained according to the principle of maximizing link gain difference for the users in the first and last user groups. The remaining user groups are added to the two existing user clusters according to the parity of the group number. The clustering should be clustered again among the users in either user cluster if the throughput summation of a user cluster in NOMA is less than that of these users in orthogonal multiple access. The simulation results show that the proposed clustering scheme can increase the system throughput by about 8% compared with the hybrid clustering scheme when the number of users requiring service is 12.

Random group recommendation model based on fuzzy clustering