Toward Fair Recommendation in Two-sided Platforms

Many online platforms today (such as Amazon, Netflix, Spotify, LinkedIn, and AirBnB) can be thought of as two-sided markets with producers and customers of goods and services. Traditionally, recommendation services in these platforms have focused on maximizing customer satisfaction by tailoring the results according to the personalized preferences of individual customers. However, our investigation reinforces the fact that such customer-centric design of these services may lead to unfair distribution of exposure to the producers, which may adversely impact their well-being. However, a pure producer-centric design might become unfair to the customers. As more and more people are depending on such platforms to earn a living, it is important to ensure fairness to both producers and customers. In this work, by mapping a fair personalized recommendation problem to a constrained version of the problem of fairly allocating indivisible goods, we propose to provide fairness guarantees for both sides. Formally, our proposed FairRec algorithm guarantees Maxi-Min Share of exposure for the producers, and Envy-Free up to One Item fairness for the customers. Extensive evaluations over multiple real-world datasets show the effectiveness of FairRec in ensuring two-sided fairness while incurring a marginal loss in overall recommendation quality. Finally, we present a modification of FairRec (named as FairRecPlus ) that at the cost of additional computation time, improves the recommendation performance for the customers, while maintaining the same fairness guarantees.

Global Multi-Source Information Fusion Management and Deep Learning Optimization for Tourism

The purpose is to solve the problems of sparse data information, low recommendation precision and recall rate and cold start of the current tourism personalized recommendation system. First, a context based personalized recommendation model (CPRM) is established by using the labeled-LDA (Labeled Latent Dirichlet Allocation) algorithm. The precision and recall of interest point recommendation are improved by mining the context information in unstructured text. Then, the interest point recommendation framework based on convolutional neural network (IPRC) is established. The semantic and emotional information in the comment text is extracted to identify user preferences, and the score of interest points in the target location is predicted combined with the influence factors of geographical location. Finally, real datasets are adopted to evaluate the recommendation precision and recall of the above two models and their performance of solving the cold start problem.

Exploiting Group Information for Personalized Recommendation with Graph Neural Networks

Personalized recommendation has become more and more important for users to quickly find relevant items. The key issue of the recommender system is how to model user preferences. Previous work mostly employed user historical data to learn users’ preferences, but faced with the data sparsity problem. The prevalence of online social networks promotes increasing online discussion groups, and users in the same group often have similar interests and preferences. Therefore, it is necessary to integrate group information for personalized recommendation. The existing work on group-information-enhanced recommender systems mainly relies on the item information related to the group, which is not expressive enough to capture the complicated preference dependency relationships between group users and the target user. In this article, we solve the problem with the graph neural networks. Specifically, the relationship between users and items, the item preferences of groups, and the groups that users participate in are constructed as bipartite graphs, respectively, and the user preferences for items are learned end to end through the graph neural network. The experimental results on the Last.fm and Douban Movie datasets show that considering group preferences can improve the recommendation performance and demonstrate the superiority on sparse users compared

Neural Matrix Factorization Recommendation for User Preference Prediction Based on Explicit and Implicit Feedback

Explicit feedback and implicit feedback are two important types of heterogeneous data for constructing a recommendation system. The combination of the two can effectively improve the performance of the recommendation system. However, most of the current deep learning recommendation models fail to fully exploit the complementary advantages of two types of data combined and usually only use binary implicit feedback data. Thus, this paper proposes a neural matrix factorization recommendation algorithm (EINMF) based on explicit-implicit feedback. First, neural network is used to learn nonlinear feature of explicit-implicit feedback of user-item interaction. Second, combined with the traditional matrix factorization, explicit feedback is used to accurately reflect the explicit preference and the potential preferences of users to build a recommendation model; a new loss function is designed based on explicit-implicit feedback to obtain the best parameters through the neural network training to predict the preference of users for items; finally, according to prediction results, personalized recommendation list is pushed to the user. The feasibility, validity, and robustness are fully demonstrated in comparison with multiple baseline models on two real datasets.

What Makes Consumers’ Intention to Purchase Paid Stickers in Personal Messenger? The Role of Personality and Motivational Factors

Many messengers and social networking services (SNSs) use emojis and stickers as a means of communication. Stickers express individual emotions well, allowing long texts to be replaced with small pictures. As the use of stickers increased, stickers were commercialized on a few platforms and showed remarkable growth as people bought and used stickers with their favorite characters, products, or entertainers online. Depending on their personality, individuals have different motivations for using stickers that determine the usefulness and enjoyment of stickers, affecting their purchase decisions. In the present study, participants (n = 302) who were randomly recruited from a university completed an online questionnaire assessing the Big Five personality characteristics, motivations for using stickers, and the technology acceptance model (TAM). Results using partial least squares structural equation modeling (PLS-SEM) revealed that each personality trait affected different motivations for using stickers. Moreover, motivations for using stickers also influenced different technology acceptance variables. Finally, perceived usefulness, enjoyment, and ease of use had a positive effect on the intention to purchase stickers. This study has implications in that it is an exploratory approach to the intention to purchase stickers, which has been investigated by few prior studies, and it sheds light on the relationship between personality, motivation, and TAM in purchasing stickers. It also suggests that personality and motivation factors can be considered in personalized recommendation services.

Design of a Personalized Recommendation System for Learning Resources based on Collaborative Filtering

At present, personalized recommendation system has become an indispensable technology in the fields of e-commerce, social network and news recommendation. However, the development of personalized recommendation system in the field of education and teaching is relatively slow with lack of corresponding application.In the era of Internet Plus, many colleges have adopted online learning platforms amidst the coronavirus (COVID-19) epidemic. Overwhelmed with online learning tasks, many college students are overload by learning resources and unable to keep orientation in learning. It is difficult for them to access interested learning resources accurately and efficiently. Therefore, the personalized recommendation of learning resources has become a research hotspot. This paper focuses on how to develop an effective personalized recommendation system for teaching resources and improve the accuracy of recommendation. Based on the data on learning behaviors of the online learning platform of our university, the authors explored the classic cold start problem of the popular collaborative filtering algorithm, and improved the algorithm based on the data features of the platform. Specifically, the data on learning behaviors were extracted and screened by knowledge graph. The screened data were combined with the collaborative filtering algorithm to recommend learning resources. Experimental results show that the improved algorithm effectively solved the loss of orientation in learning, and the similarity and accuracy of recommended learning resources surpassed 90%. Our algorithm can fully satisfy the personalized needs of students, and provide a reference solution to the personalized education service of intelligent online learning platforms.

Design of Context Cluster and Load Balanced Video Recommendation System in Cloud Computing

Nowadays, in online social networks, there is an instantaneous extension of multimedia services and there are huge offers of video contents which has hindered users to acquire their interests. To solve these problem different personalized recommendation systems had been suggested. Although, all the personalized recommendation system which have been suggested are not efficient and they have significantly retarded the video recommendation process. So to solve this difficulty, context extractor based video recommendation system on cloud has been proposed in this paper. Further to this the system has server selection technique to handle the overload program and make it balanced. This paper explains the mechanism used to minimize network overhead and recommendation process is done by considering the context details of the users, it also uses rule based process and different algorithms used to achieve the objective. The videos will be stored in the cloud and through application videos will be dumped into cloud storage by reading, coping and storing process.