Bayesian Personalized Ranking Research Articles

PGCF: Perception graph collaborative filtering for recommendation

Extensive studies have fully proved the effectiveness of collaborative filtering (CF) recommendation models based on graph convolutional networks (GCNs). As an advanced interaction encoder, however, GCN-based CF models do not differentiate neighboring nodes, which will lead to suboptimal recommendation performance. In addition, most GCN-based CF studies pay insufficient attention to the loss function and they simply select the Bayesian personalized ranking (BPR) loss function to train the model. However, we believe that the loss function is as important as the interaction encoder and deserves more attentions. To address the above issues, we propose a novel GCN-based CF model, named Perception Graph Collaborative Filtering (PGCF). Specifically, for the interaction encoder, we design a neighborhood-perception GCN to enhance the aggregation of interest-related information of the target node during the information aggregation process, while weakening the propagation of noise and irrelevant information to help the model learn better embedding representation. For the loss function, we design a margin-perception Bayesian personalized ranking (MBPR) loss function, which introduces a self-perception margin, requiring the predicted score of the user-positive sample to be greater than that of the user-negative sample, and also greater than the sum of the predicted score of the user-negative sample and the margin. The experimental results on five benchmark datasets show that PGCF is significantly superior to multiple existing CF models.

A personalized product recommendation model in e-commerce based on retrieval strategy

In recent years, online shopping is one of the routine parts in people’s life. It is convenient and takes less effort to purchase it. Regarding the increasing revolution of e-commerce businesses, recommendation engine plays a crucial role in them. Recommendation engines are very popular and easy to implement to their platform nowadays. Due to the extremely high competition of e-commerce businesses, the operation needs to integrate the recommender wisely. This study presents a comprehensive approach to improving user experience and engagement on e-commerce platforms through the implementation of an implicit personalized product recommendation engine. Collaborating with the H&M Group, the research combines the strength of each recommending algorithms which are collaborative filtering, popularity, and Bayesian personalized ranking to develop a robust recommendation system. By leveraging a retrieval strategy that combines multiple algorithmic techniques and evaluating candidates using machine learning models which comprise LightGBM and Deep Neural Network, the study achieves promising results. The authors utilize two popular technical metrics to evaluate their models which are mean average precision at K candidates (MAP@K) and mean average recall at K candidates (MAR@K). The empirical result indicates that the LightGBM model has remarkable performance than Deep Neural Network model, which are 0.06 versus 0.02 respectively in MAP@K and 0.03 versus 0.01 respectively in MAR@K when both recommending ways is at 50 items. Overall, this research contributes a novel framework that addresses the challenges of analyzing large-scale data, cold-start problems, and personalization, thereby enhancing the user experience, and driving sales on e-commerce platforms.

Dual-blend insight recommendation system for e-commerce recommendations and enhance personalization

E-commerce, short for electronic commerce, refers to the buying and selling of goods and services over the internet. This digital transaction model has revolutionized the way businesses operate and consumers shop. In response to the burgeoning complexity of e-commerce datasets, this work addresses the need for advanced recommendation systems. This work introduces the dual-blend insight recommendation system (DIRS) model for personalized e-commerce recommendation system. The DIRS model involves dataset loading, preprocessing, and feature extraction, enabling training with recurrent neural network (RNN) and Bayesian personalized ranking (BPR) models. Recommendations are generated based on user-defined functions, i.e., location and session, and evaluation metrics such as hit rate (HR) and mean reciprocal rate (MRR) highlight DIRS’s superior performance. The model is evaluated using the Tmall dataset. Results reveal DIRS consistently outperforms alternative algorithms, showcasing its effectiveness in 10k and 20k recommendation sets. This study provides valuable insights into optimizing e-commerce recommendations, emphasizing DIRS as a powerful model for enhancing user experience and engagement.

CrossGCL: Cross-pairwise graph contrastive learning for unbiased recommendation

Popularity bias is commonly observed in recommendation results. Directly fitting biased data can significantly affect the quality of recommendations for long-tail items. To eliminate popularity bias, we propose a novel recommendation method called CrossGCL. Unlike existing Bayesian Personalized Ranking (BPR) based unbiased recommendation methods, CrossGCL employs Cross Pairwise Ranking (CPR) as its primary task, in an attempt to accurately predict user preferences and promote niche items. Furthermore, to better utilize unsupervised data for debiasing, CrossGCL introduces an innovative layer-wise graph contrastive learning approach based on odd-numbered and even-numbered layers and incorporates noise augmentation to construct pretext tasks. Then CrossGCL jointly trains pretext task with primary task for Top-K recommendation. Experimental results on MovieLens-10M, Netflix, and Alibaba-iFashion demonstrate that, compared to baselines, CrossGCL effectively reduces the proportion of popular items in recommendation lists while improving recommendation accuracy. The code is available at https://github.com/RYoto116/CrossGCL.

SiReN: Sign-Aware Recommendation Using Graph Neural Networks.

In recent years, many recommender systems using network embedding (NE) such as graph neural networks (GNNs) have been extensively studied in the sense of improving recommendation accuracy. However, such attempts have focused mostly on utilizing only the information of positive user-item interactions with high ratings. Thus, there is a challenge on how to make use of low rating scores for representing users' preferences since low ratings can be still informative in designing NE-based recommender systems. In this study, we present SiReN, a new Sign-aware Recommender system based on GNN models. Specifically, SiReN has three key components: 1) constructing a signed bipartite graph for more precisely representing users' preferences, which is split into two edge-disjoint graphs with positive and negative edges each; 2) generating two embeddings for the partitioned graphs with positive and negative edges via a GNN model and a multilayer perceptron (MLP), respectively, and then using an attention model to obtain the final embeddings; and 3) establishing a sign-aware Bayesian personalized ranking (BPR) loss function in the process of optimization. Through comprehensive experiments, we empirically demonstrate that SiReN consistently outperforms state-of-the-art NE-aided recommendation methods.

ABNS: Association-based negative sampling for collaborative filtering

The purpose of the recommendation systems (RS) is to provide personalized information filtering for users by analyzing their interactions with items. However, the sparsity of data in RS, which means that the number of negative samples is much larger than the number of positive samples, causes the necessity of the negative sampling method. Previous work has shown that an appropriate negative sampling method, rather than simple random negative sampling, can significantly improve the performance of recommendation algorithms. Unfortunately, the existing works have significant shortcomings in explainability and computational complexity. To remedy these shortcomings, we conducted a theoretical analysis of the gradient of the Bayesian Personalized Ranking (BPR) loss function and propose an explainable negative sampling method. Moreover, we conducted experiments on real-world datasets and find that the proposed negative sampling method improves the recommendation performance by an average of 10.23% over uniform negative sampling. Finally, the effectiveness of the proposed method is demonstrated through comparison with other methods and hyperparameter sensitivity analysis.

Friends Recommendation on Social Networks using the Bayesian Personalized Ranking-Matrix Factorization

In the digital landscape of social networking, the challenge of improving friend recommendation systems is pivotal for enhancing user interaction and fostering social connections. Addressing this challenge, the current study innovates by fusing Bayesian Personalized Ranking (BPR) with Matrix Factorization (MF), culminating in a novel BPR-MF model designed for the intricacies of social network relationships. The study harnesses a rich dataset from LastFM, comprising 27,806 interactions among 7,624 users, to analyze mutual follower patterns and augment the precision of friend recommendations. Through rigorous preprocessing and systematic evaluation of the BPR-MF model against different numbers of latent factors, the research uncovers that a configuration of 20 latent factors is most effective, achieving an RMSE of 0.156 and an AUC ROC of 0.800. This discovery addresses the critical problem of balancing computational complexity with prediction accuracy in recommendation models. It also demonstrates the necessity for a nuanced, data-driven approach to generate relevant social connections. The research sets a new direction for future studies aiming to capitalize on user interaction data to offer precise friend suggestions, all while upholding user privacy and avoiding reliance on personal data.

WeightedSLIM: A Novel Item-Weights Enriched Baseline Recommendation Model

This paper proposes a novel weight-enriched ranking-based baseline model, WeightedSLIM, aiming to provide more accurate top-N recommendations from implicit user feedback. The model utilizes ItemRank to calculate the ranking score of each item, which is then used as an item weight within the Sparse Linear Model (SLIM), while using Bayesian Personalized Ranking (BPR) to optimize the item similarity matrix. Experiments, conducted for performance comparison of the proposed model with existing recommendation models, demonstrate that it can indeed provide better recommendations and can be used as a strong baseline for top-N recommendations.

Neural_BPR: Multi-processing popularity bias mitigating method in recommendation systems

E-commerce recommendations address the problem of information overload, but recent research has identified the phenomenon of popularity bias in recommendation mechanisms. This phenomenon tends to homogenise recommendation results, i.e. popular products are widely recommended and non-popular products are under-exposed. The persistent presence of popularity bias in recommendation systems can lead to several negative consequences: (1) Consumers are unable to attain satisfactory personalized shopping experiences. (2) Small and medium enterprises are deprived of fair competitive opportunities, making survival challenging. (3) The user base of e-commerce platforms dwindles, leading to reduced transaction volumes. In this paper, we propose a multi-process fusion debiasing method. Convolutional neural networks are used to extract feature information to be embedded in the recommendation algorithm for pre-processing, in-processing and post-processing in turn to mitigate popularity bias and simultaneously enhance recommendation utility. The effectiveness of the proposed debiasing method is validated using the Bayesian Personalized Ranking (BPR) recommendation model as an example. Experimental results on two public datasets demonstrate that our model outperforms two baseline models and a state-of-the-art model. Specifically, compared to the traditional BPR model, the alleviation of popularity bias is around 70% to 80%, and the recommendation utility increases by approximately 30% to 50%.

BayesSentiRS: Bayesian sentiment analysis for addressing cold start and sparsity in ranking-based recommender systems

Recommendation systems are widely used to filter massive information. However, they often face the challenges of cold start and sparsity problems, limiting their effectiveness. Bayesian Personalized Ranking (BPR), which focuses on predicting the relative order of user items, has been conventionally proposed to address these challenges for ranking-based recommendation systems. However, conventional BPR approaches rely solely on implicit feedback and lack semantic and visual sentiment information. In this study, we propose a novel multi-modal BPR method that incorporates both semantic and visual sentiment to capture more nuanced user preferences and provide more accurate recommendations. Experimental evaluations show that the performance of our proposed method across seven metrics outperforms conventional BPR in both the review-rich and image-rich scenarios, indicating the potential and significance of considering sentiment for improving the performance of BPR recommendation systems.

Personalized recommendation with hybrid feedback by refining implicit data

In personalized recommender systems, the collaborative filtering (CF) recommendation approaches have been widely used to predict the preferences of users in real-world applications. Among them, Bayesian personalized ranking (BPR) attracts much attention as it can easily explore the binary form of implicit feedback. However, it still suffers from the absence problem of negative feedback. To address this issue, this paper proposes a hybrid-feedback collaborative filtering model by jointly exploiting the explicit and implicit feedback. Based on the assumption that users prefer items with high ratings, this work firstly introduces the definition of explicit rating data to the BPR model and further proposes an improved Bayesian personalized ranking (IBPR) model to jointly extract the implicit feedback features of users and items. The IBPR model alleviates the problem of lack of negative feedback and promotes the anti-noise performance of the recommender system. Then the IBPR and BiasSVD (Biased Singular Value Decomposition) models are combined to further extract explicit latent features of users as well as items and construct the hybrid-feedback CF model. In this model, the user–item ranking matrix is reconstructed based on the extracted implicit feedback features, and the rating matrix is constructed based on the extracted explicit feedback features. Our proposed method is evaluated on five public datasets and achieves the competitive performance.

Preference-aware Bayesian Personalized Ranking for Point-of-interest recommendation

Point-of-interest (POI) recommendation has become one of the research highlights in the field of recommender systems due to the prosperity of location-based social networks in recent years. Various techniques have been proposed to improve the performance of the personalized recommendation service. Embedding-based methods have shown promising effect and attracted great attention for their flexibility and efficiency. Bayesian Personalized Ranking (BPR), as a famous optimization algorithm, has been widely used to learning the parameters of Embedding-based models in the recommendation scenario. However, existing Bayesian Personalized Ranking and its follow-up methods ignore the unique user preference when constructing the positive and negative samples, leading a suboptimal performance. To overcome this limitation, we propose a novel method named preference-aware Bayesian Personalized Ranking (PABPR) according to empirical analyses on real-world datasets. The empirical analyses show that a user tend to visit a POI with categories which have been visited before. Thus, the key idea of PABPR is to introduce such user behaviors into the sample constructing process. PABPR is a general method which could be used for training various Embedding methods. Extensive experiments show that PABPR can lead a superior model performance compare to BPR and its variant methods.

Learning Peer Recommendation Based on Weighted Heterogeneous Information Networks on Online Learning Platforms

With the development of online education, there is an urgent need to solve the problem of the low completion rate of online learning courses. Although learning peer recommendation can effectively address this problem, prior studies of learning peer-recommendation methods extract only a portion of the interaction information and fail to take into account the heterogeneity of the various types of objects (e.g., students, teachers, videos, exercises, and knowledge points). To better motivate students to complete online learning courses, we propose a novel method to recommend learning peers based on a weighted heterogeneous information network. First, we integrate the above different objects, various relationships between objects, and the attribute values to links in a weighted heterogeneous information network. Second, we propose a method for automatically generating all meaningful weighted meta-paths to extract and identify meaningful meta-paths. Finally, we use the Bayesian Personalized Ranking (BPR) optimization framework to discover the personalized weights of target students on different meaningful weighted meta-paths. We conducted experiments using three real datasets, and the experimental results demonstrate the effectiveness and interpretability of the proposed method.

Read between the interactions: Understanding non-interacted items for accurate multimedia recommendation

This paper addresses the problem of multimedia recommendation that additionally utilizes multimedia data, such as visual and textual modalities of items along with the user-item interaction information. Existing multimedia recommender systems assume that all the non-interacted items of a user have the same degree of negativity, thus regarding them as candidates for negative samples when training the model. However, this paper claims that a user?s non-interacted items do not have the same degree of negativity. We classify these non-interacted items of a user into two kinds of items with different characteristics: unknown and uninteresting items. Then, we propose a novel negative sampling technique that only considers the uninteresting items (i.e., rather than the unknown items) as candidates for negative samples. In addition, we show that using the multiple Bayesian personalized ranking (BPR) losses with both unknown and uninteresting items (i.e., all the noninteracted items) in existing multimedia recommendation methods is very effective in improving recommendation accuracy. By conducting extensive experiments with three real-world datasets, we show the superiority of our ideas. Our ideas can be easily and orthogonally applied to any multimedia recommender systems.

Ranking-based contrastive loss for recommendation systems

The recommendation system is fundamental technology of the internet industry intended to solve the information overload problem in the big data era. Top-k recommendation is an important task in this field. It generally functions through the comparison of positive pairs and negative pairs based on Bayesian personalized ranking (BPR) loss. We find that the contrastive loss (CL) function used in contrastive learning is well-suited for top-k recommendation. However, there are two problems in the existing loss functions. First, all samples are treated the same, and hard samples are not considered. Second, all nonpositive samples are considered negative samples, which ignores the fact that they are unlabelled data containing items that users may like. Moreover, in our experiments, we find that when items are sorted by their similarities to the user, many negative items (or samples) appear before the positive items. We regard these negative items as hard samples and those at the top as potentially positive samples due to their high level of similarities with users. Therefore, we propose a ranking-based contrastive loss (RCL) function to exploit both hard samples and potentially positive samples. Experimental results demonstrate the effectiveness, broad applicability, and high training efficiency of the proposed RCL function. The code and data are available at https://github.com/haotangxjtu/RCL.

MSBPR: A multi-pairwise preference and similarity based Bayesian personalized ranking method for recommendation

For addressing the “One-Class Collaborative Filtering” (OCCF) problem in recommendation systems, in which the obtained user information is all single-type positive feedback, the current mainstream methods are all based on the idea of pairwise preference learning. The Bayesian Personalized Ranking (BPR) method is a classical representative of such an idea. However, the assumption in BPR, that “users tend to prefer items that they have once interacted”, may not always hold in reality. This is because for non-interacted items, a user may have different perspectives, such as potentially favorite, dislike, or something in between. For mitigating the above-mentioned issue, this paper proposes a Multi-pairwise preference and Similarity based BPR method, termed as MSBPR for brevity. MSBPR utilizes additional auxiliary feedback information to excavate and infer deeper-level user preferences. Subsequently, MSBPR borrows ideas from traditional item/user-based collaborative filtering methods to further divide non-interacted items from the angle of item/user, respectively. Afterwards, MSBPR constructs four preferences on top of the divided items, and accordingly builds up the multiple pairwise preference assumption. To optimize MSBPR, we derive an efficient learning algorithm based on the stochastic gradient descent algorithm. The computational complexity of MSBPR is also theoretically analyzed. Comprehensive experimental results demonstrate the effectiveness and efficiency of MSBPR over ten state-of-the-art methods on six benchmark real-world datasets.

Bayesian Personalized Ranking with the Synthesis of Multiple User and Item Classification

<p>In recent years, rapidly evolving networks have facilitated data interactions between users and goods, and recommender systems have emerged due to the needs of the times. Implicit feedback in recommender systems always is divided into the observed item set and unobserved item set. However, items in the unobserved set often are treated equally, and the items that are of potential interest to users are not sufficiently exploited from unobserved items. In this paper, we come up with a method to solve the problem of the insufficient exploration of unrated items, and we contribute to the additional alleviation of rating sparseness. To do so, we propose a novel Bayesian personalized ranking with the synthesis of multiple users and item classification (BPRS). For each user, we divide the items into three categories, which are positive, interest and negative. We conduct multiple user classification and item classification to exploit the items that are of interest and generate the final ranking. Experiments on three real-world datasets demonstrate the effectiveness of our algorithm for greatly improving the accuracy of recommendation results and alleviating the cold-start problem.</p> <p> </p>

A Survey of One Class E-Commerce Recommendation System Techniques

Although several recommendation algorithms are widely used by both commercial and non-commercial platforms, they face unique challenges such as sparse data sets and the absence of negative or “neutral” feedback. One-class algorithms attempt to overcome the data sparsity problem by using the implicit feedback inherent in user’s clicks and purchases, which are deduced from both positive and negative feedback. Existing literature uses several heuristic strategies to derive the negative samples needed for training, such as using random sampling or utilizing user-item interaction. However, these assumptions do not always reflect reality. In addition, with the explosive increase in the availability of big data for training recommendation systems, these methods might not adequately encapsulate the representations of the latent vectors. In this paper, we address the common issues of one-class recommendation and provide a survey on approaches that have been used to mitigate the existing challenges. To tackle the identified problems, we propose a neural network-based Bayesian Personalized Ranking (BPR) for item recommendation and personalized ranking from implicit feedback. BPR provides an optimization criterion derived from Bayesian analysis of a problem to develop an optimized model for such a problem. We conduct several experiments on two varieties of MovieLens datasets to illustrate the performance of the proposed approach. Our approach shows an impressive result in mitigating the issues of one-class recommendation when compared with the complexity of the state-of-the-art methods.

A framework for unbiased explainable pairwise ranking for recommendation

<h2>Abstract</h2> Recent research in recommender systems has demonstrated the advantages of pairwise ranking in recommendation. In this work, we focus on the state-of-the-art pairwise ranking loss function, Bayesian Personalized Ranking (BPR), and aim to address two of its limitations, namely: (1) the lack of explainability and (2) exposure bias. We propose a recommendation framework that encompasses various loss functions that are based on BPR and which aim to mitigate the aforementioned limitations. Our open-source framework includes code to train and tune state-of-the-art pairwise ranking recommender systems on benchmark datasets and evaluate them based on the three criteria of ranking accuracy, explainability, and popularity debiasing.

SPR: Similarity pairwise ranking for personalized recommendation

Bayesian personalized ranking (BPR) has been proposed as an effective method to model pairwise learning, and it is widely used in many personalized recommender systems. However, the effectiveness of BPR can be seriously affected by an imbalanced data distribution because it tends to rank popular items ahead of personalized items. As a result, the personalized needs of users cannot be well met. In this paper, we propose a novel personalized recommendation method called similarity pairwise ranking (SPR) to rank users’ favorite items first. SPR eliminates the differences in the scores between popular and personalized items based on their similarity by using a new penalty. In such a way, the SPR-enhanced recommendation will render meaningful and personalized results that better meet the individual needs of users, and it overcomes the negative impact of imbalanced datasets. We design a model to illustrate the improvement of SPR: similarity pairwise ranking matrix factorization (SPRMF). Experimental results obtained using six datasets indicate the superiority in recommendation quality of SPRMF over the recent state-of-the-art methods.