Direction-Aware User Recommendation Based on Asymmetric Network Embedding

User recommendation aims at recommending users with potential interests in the social network. Previous works have mainly focused on the undirected social networks with symmetric relationship such as friendship, whereas recent advances have been made on the asymmetric relationship such as the following and followed by relationship. Among the few existing direction-aware user recommendation methods, the random walk strategy has been widely adopted to extract the asymmetric proximity between users. However, according to our analysis on real-world directed social networks, we argue that the asymmetric proximity captured by existing random walk based methods are insufficient due to the inbalance in-degree and out-degree of nodes. To tackle this challenge, we propose InfoWalk, a novel informative walk strategy to efficiently capture the asymmetric proximity solely based on random walks. By transferring the direction information into the weights of each step, InfoWalk is able to overcome the limitation of edges while simultaneously maintain both the direction and proximity. Based on the asymmetric proximity captured by InfoWalk, we further propose the qualitative (DNE-L) and quantitative (DNE-T) directed network embedding methods, capable of preserving the two properties in the embedding space. Extensive experiments conducted on six real-world benchmark datasets demonstrate the superiority of the proposed DNE model over several state-of-the-art approaches in various tasks.

User recommendation in online health communities using adapted matrix factorization

PurposeOnline health communities (OHCs) are platforms that help health consumers to communicate with each other and obtain social support for better healthcare outcomes. However, it is usually difficult for community members to efficiently find appropriate peers for social support exchange due to the tremendous volume of users and their generated content. Most of the existing user recommendation systems fail to effectively utilize the rich social information in social media, which can lead to unsatisfactory recommendation performance. The purpose of this study is to propose a novel user recommendation method for OHCs to fill this research gap.Design/methodology/approachThis study proposed a user recommendation method that utilized the adapted matrix factorization (MF) model. The implicit user behavior networks and the user influence relationship (UIR) network were constructed using the various social information found in OHCs, including user-generated content (UGC), user profiles and user interaction records. An experiment was conducted to evaluate the effectiveness of the proposed approach based on a dataset collected from a famous online health community.FindingsThe experimental results demonstrated that the proposed method outperformed all baseline models in user recommendation using the collected dataset. The incorporation of social information from OHCs can significantly improve the performance of the proposed recommender system.Practical implicationsThis study can help users build valuable social connections efficiently, enhance communication among community members, and potentially contribute to the sustainable prosperity of OHCs.Originality/valueThis study introduces the construction of the UIR network in OHCs by integrating various social information. The conventional MF model is adapted by integrating the constructed UIR network for user recommendation.

Humanized Recommender Systems: State-of-the-art and Research Issues

Psychological factors such as personality, emotions, social connections , and decision biases can significantly affect the outcome of a decision process. These factors are also prevalent in the existing literature related to the inclusion of psychological aspects in recommender system development. Personality and emotions of users have strong connections with their interests and decision-making behavior. Hence, integrating these factors into recommender systems can help to better predict users’ item preferences and increase the satisfaction with recommended items. In scenarios where decisions are made by groups (e.g., selecting a tourism destination to visit with friends), group composition and social connections among group members can affect the outcome of a group decision. Decision biases often occur in a recommendation process, since users usually apply heuristics when making a decision. These biases can result in low-quality decisions. In this article, we provide a rigorous review of existing research on the influence of the mentioned psychological factors on recommender systems. These factors are not only considered in single-user recommendation scenarios but, importantly, also in group recommendation ones, where groups of users are involved in a decision-making process. We include working examples to provide a deeper understanding of how to take into account these factors in recommendation processes. The provided examples go beyond single-user recommendation scenarios by also considering specific aspects of group recommendation settings.

Research on the Application of User Recommendation Based on the Fusion Method of Spatially Complex Location Similarity

Since the user recommendation complex matrix is characterized by strong sparsity, it is difficult to correctly recommend relevant services for users by using the recommendation method based on location and collaborative filtering. The similarity measure between users is low. This paper proposes a fusion method based on KL divergence and cosine similarity. KL divergence and cosine similarity have advantages by comparing three similar metrics at different K values. Using the fusion method of the two, the user’s similarity with the preference is reused. By comparing the location-based collaborative filtering (LCF) algorithm, user-based collaborative filtering (UCF) algorithm, and user recommendation algorithm (F2F), the proposed method has the preparation rate, recall rate, and experimental effect advantage. In different median values, the proposed method also has an advantage in experimental results.

A User Recommendation Model for Answering Questions on Brainly Platform

Brainly is a Community Question Answer (CQA) application that allows students or parents to ask questions related to their homework. The current mechanism is that users ask questions, then other users who are in the same subject interest can see and answer it. As a reward for answering questions, Brainly gives points. The number of points varies by question. The greater of total points users have, Brainly will automatically display them in the smartest user leaderboard on the site's front page. But sometimes, some users do not have good activity in answering questions. Thus, it is possible to have an urgent question that has not been answered by anyone. This study implements Fuzzy C-Means cluster method to improve Brainly's feature regarding the speed and accuracy of answers. The idea is to create student clusters by utilizing the smartest students' leaderboard, subjects interest, and answering activities. The stages applied in this research started with Data Extraction, Preprocessing, Cluster Process, and User Recommender. The optimal number of clusters in the answerer recommendation in the Brainly platform is 2 clusters. The value of the fuzzy partition coefficient for two clusters reached 0.97 for Mathematics and 0.93 for Indonesian. Meanwhile, the results of the recommendations were influenced by answer ratings. Many numbers of the answer are not given rating because the possibility of the answers are not appropriate or user's insensitivity in giving ratings.

User Recommendation for Data Sharing in Social Internet of Things

As various types of data are generated on the social Internet of things (SIoT), which combine the Internet of things (IoT) and social networks, the relations of IoT devices should be established for necessary data exchange. In this paper, we propose a user recommendation scheme that facilitates data sharing through an analysis of an interaction between an IoT device and a user in the SIoT. An interrelation between a user and an IoT device as well as an interrelation between users exist simultaneously in the SIoT. Hence, the interaction between users must be analyzed to identify the interest keywords, and the interaction between IoT devices and users to determine the user’s preference of IoT device. Moreover, the proposed scheme calculates the similarity between users based on the IoT device preference based on IoT device usage frequency and interest keywords, which are identified through an analysis between the user and IoT device and that between users. Subsequently, it recommends top-N users who have a high similarity as the users for data sharing. Furthermore, the performance of the proposed scheme is verified through performance evaluation based on the precision, recall, and F-measure.