Traditional Recommender Systems Research Articles

The Application of Personalized Recommendation System in the Cross-Regional Promotion of Provincial Intangible Cultural Heritage

With the accelerated trend of globalization and modernization, many provincial intangible cultural heritages (PICH) are in danger of being lost. In the context of Internet technology, the use of digital multimedia for personalized recommendation is an effective way to promote the transmission of intangible cultural heritage. However, traditional recommendation systems tend to treat different members as homogeneous objects, ignoring the relationship between members’ professional backgrounds and the inherent properties of items, and cannot truly solve the problem of conflicting preferences in the integration process. In view of this, this paper proposes a group recommendation system based on nonnegative matrix decomposition. First, the group rating information is decomposed into a user matrix and item matrix by nonnegative matrix decomposition. Then, the item affiliation matrix and member expertise matrix are calculated by using the affiliation and expertise weights for the two matrices, respectively, and the contribution degree of each member to different item ratings is obtained from them. Finally, the group preference model is constructed by weighted fusion of members’ preferences based on their contribution degrees, and different recommendation lists are generated for different user preferences. The experimental results prove that this system has high recommendation accuracy in cross-regional promotion of PICH.

Movie Recommendation Systems Using Actor-Based Matrix Computations in South Korea

A recommendation system saves the user the trouble of searching for information and analyzes their profile to recommend the most suitable content. A variety of techniques, including content-based, collaborative, and knowledge-based techniques, have been proposed for performing recommendations. Recommendation systems are employed to suggest content, such as books, music, and videos; furthermore, they are often used in e-commerce. In particular, the South Korean film industry recommends movies using a collaborative filtering method based on genres, which is commonly used in film recommendation systems. However, this method can be ineffective when users initially encounter film recommendation services or have specific preferences with respect to movies, such as preferences regarding actors or directors. This motivated us to propose an actor-based recommendation system using the content-based filtering that considers actors’ filmography information and the genres of 509 South Korean movies. The effectiveness and performance of the suggested system are evaluated in comparison with those of a traditional genre-oriented recommendation system.

Trust-aware denoising autoencoder with spatial-temporal activity for cross-domain personalized recommendations

Recently, cross-domain recommendation systems have been very helpful in improving the quality of recommendation and solving the problem of cold start and data sparsity. Cross-domain recommender systems allow the transfer of knowledge from one domain with dense ratings to other domain with sparse ratings. Such transfer of knowledge helps in addressing the data sparseness and cold start issues in traditional recommender systems. Although cross-domain recommendations have evolved significantly, yet employment of various factors, such as time, trust, and location remains a challenge. Most of the existing approaches ignore the important fact that at what specific time the user may be interested in the recommended item. Moreover, a person’s trust level and sentiments may be influenced by the variation in the location and time, thereby affecting the decision making. In this paper, we propose a cross-domain recommender system that not only takes into account the time at finer granularity levels (e.g., hours, days, weeks, etc.), but also considers a persons location, trust level, and sentiment analysis while computing recommendations. Our proposed model, named as, Trust-Aware Spatial-Temporal Activity based Denoising Autoencoder (TSTDAE), employs autoencoder-based deep-learning models to generate top-N recommendations for a given user and addresses the cold-start problem in the cross-domain scenario of ‘User Overlap’. The proposed work is fivefold: i) Filter out the users’ biased profiles based on sentiment analysis. ii) Learn the features using autoencoder and then perform clustering among the users of source and target domains to discover the best neighbors. iii) Compute the trust and ratio of preference bias between active user (the user to whom top-N items are recommended) and their neighbors and grade the neighbors based on unbiased preferences iv) Project the best time for recommending the items to an active user and generate the top-N recommendations. We have evaluated the proposed model on a public dataset of e-commerce retail service ‘AliExpress’ for the evaluation metrics: Precision, Mean Absolute Precision (MAP), Normalized Discounted Cumulative Gain (NDCG), and Hit Ratio (HR). The experimental results showed improved performance of the proposed system over the existing models.

Remembering past and predicting future: a hybrid recurrent neural network based recommender system.

Traditional recommender systems (RS) assume users’ taste to be static (taste remains same over time) and reactive (a change in taste cannot be predicted and is observed only after it occurs). Further, traditional RS restricts the recommendation process to candidate items generation. This work aims to explore two phases of RS, i.e., Candidate Generation as well as Candidate Ranking. We propose a RS from a multi-objective (short-term prediction, long-term prediction, diversity, and popularity bias) perspective which was previously overlooked. The sequential and non-sequential behavior of users is exploited to predict future behavioral trajectories with the consideration of short-term and long-term prediction using recurrent neural networks and nearest neighbors approach. Further, a novel candidate ranking method is introduced to prevent users from being entangled in recommended items. On multiple datasets, largest being MovieLens (ML) 1M, our model shows excellent results achieving a hit rate and short-term prediction success of 58% and 71% respectively on ML 1M. Further, it implicitly handles two important parameters, i.e., diversity and item popularity with a success rate of 59.22% and 34.28% respectively.

Deep transfer learning with multimodal embedding to tackle cold-start and sparsity issues in recommendation system.

Recommender systems (RSs) have become increasingly vital in the modern information era and connected economy. They play a key role in business operations by generating personalized suggestions and minimizing information overload. However, the performance of traditional RSs is limited by data sparseness and cold-start issues. Though deep learning-based recommender systems (DLRSs) are very popular, they underperform when considering rating matrices with sparse entries. Despite their performance improvements, DLRSs also suffer from data sparsity, cold start, serendipity, and generalizability issues. We propose a multistage model that uses multimodal data embedding and deep transfer learning for effective and personalized product recommendations, and is designed to overcome data sparsity and cold-start issues. The proposed model includes two phases. In the first—offline—phase, a deep learning technique is implemented to learn hidden features from a large image dataset (targeting new item cold start), and a multimodal data embedding is used to produce dense user feature and item feature vectors (targeting user cold start). This phase produces three different similarity matrices that are used as inputs for the second—online—phase to generate a list of top-n relevant items for a target user. We analyzed the accuracy and effectiveness of the proposed model against the existing baseline RSs using a Brazilian E-commerce dataset. The results show that our model scored 0.5882 for MAE and 0.4011 for RMSE which is lower than baseline RSs which indicates that the model achieved an improved accuracy and was able to minimize the typical cold start and data sparseness issues during the recommendation process.

A hybrid recommendation technique using topic embedding for rating prediction and to handle cold-start problem

Recommender systems aim to estimate item ratings and recommend items based on the users’ interests. The traditional recommender systems generally consider user–item rating information for rating prediction, but they suffer from various limitations, such as data sparsity, black-box recommendation, and cold-start problems. As a result, researchers have proposed amalgamating contextual information with rating data to provide effective recommendations. Although user-generated data in the form of reviews are a rich source of contextual information, they are rarely utilized in recommender algorithms. This study presents a hybrid recommendation technique, called RecTE, using rating data and topic embedding, which is an amalgamation of word embedding and topic modeling techniques. The novelty of RecTE lies in predicting item ratings using topic embeddings learned by incorporating local and global contextual information and integrating them with user-based collaborative filtering. RecTE is empirically evaluated over three real-world datasets – YelpNYC, YelpZip and TripAdvisor. This technique performs significantly better in comparison to nine baselines and five state-of-the-art recommendation techniques. On empirical analysis, we found that incorporating topic embedding in RecTE makes it capable of performing significantly better and handle cold-start problems effectively in comparison to the existing recommendation approaches.

A content-based recommender system with consideration of repeat purchase behavior

With the increasing popularity of online shopping and information explosion, personalized recommender systems for e-commerce become more and more necessary, which helps customers find the desired products efficiently among variety of categories based on their previous behavior such as buying pattern and rating history. However, most recommender systems for e-commerce adopt binary (purchase/non-purchase) or subjective weighting methods to represent the customer preferences, which is hard to predict their profiles precisely since rapid change in tastes. Therefore, this study focuses on the application of transactional data. A personalized recommender system for e-commerce (PROSE) is proposed in order to enhance the quality of recommendations by integrating the architecture of traditional content-based recommender system with a new component called feedback adjuster, which is designed to make customer implicit feedback reflects the reality of preferences as possible through taking into consideration their behavior of repeat purchase. The computational results indicate that the proposed algorithm is able to outperform other algorithms.

Federated low-rank tensor projections for sequential recommendation

Recommender systems have achieved great success in many fields, benefiting from the accessibility of massive amounts of user behavior data. However, the increasing concern about privacy issues brings new research challenges to traditional data-centralized recommender systems. As a privacy-preserving machine learning technique, federated learning provides potential solutions to these research challenges, by jointly training perceptive global recommendation models based on the decentralized user behavior data. Although existing federated recommendation methods have achieved promising results, they still suffer the following limitations. Firstly, they mainly focus on the horizontal federated recommendation scenarios where the participants are individual users, without considering the scenarios where participants are organizations. Secondly, they tend to model static user preferences and cannot learn the preference evolution in users’ interaction sequences. To address these limitations, we propose a horizontal Federated recommendation framework for Sequential Recommendation (FedSeqRec), where the participants are organizations. As the data in different organizations are usually not independent and identically distributed (Non-IID), federated training tends to yield suboptimal results. To alleviate this problem, we propose a federated re-average algorithm, which takes the consistency of the client’s data distribution with the overall data distribution into consideration. Moreover, we also leverage a low-rank tensor projection to model user long-term preferences. Experimental results on two real-world datasets demonstrate that FedSeqRec outperforms state-of-the-art federated recommendation methods. The implementation code of FedSeqRec is available at https://github.com/MuziLee-x/FedSeqRec.

A Reliable and Lightweight Trust Inference Model for Service Recommendation in SIoT

In the era of Internet of Things (IoT), millions of heterogeneous IoT devices generate an explosion of data and services waiting to be discovered. The convergence of IoT with social networks (SIoT) interconnects multiple IoT applications and alleviates the common data sparsity and cold start problems in traditional recommendation systems. However, the social trust relationships may also be very sparse, which affects the accuracy of trust-based recommendation systems. Meanwhile, mobile devices have limited resources and are more vulnerable to malicious attacks in the IoT environment. In order to complete the trust relationship and further improve the trust-based recommendation performance, we propose a reliable and lightweight trust inference model for service recommendation in SIoT, called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TIRec</i> . First, we obtain a comprehensive weighted centrality metric (LGWC) considering both local and global contexts. Based on this, we propose a corresponding lightweight trust path selection algorithm. Then, we present a reliable trust inference calculation algorithm consist of trust propagation and aggregation strategy, which can efficiently resist two common malicious attacks. Finally, we incorporate the rating, direct trust, and indirect trust together into the matrix factorization model, and integrate the influence of truster and trustee to obtain the synthetic model for rating predication. To the best of our knowledge, this article is the first to integrate trust inference algorithm into the trust-based recommendation systems. The extensive experiments are conducted on three real-world data sets, and the results show that our <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TIRec</i> model performs better than other advanced recommendation models in both “all users” view and “cold start users” view.

Does Utilizing Online Social Relations Improve the Diversity of Personalized Recommendations?

Personalized recommendations are widely used to improve customer experience and drive sales. Traditional recommender systems typically focus on using accuracy as the key metric to evaluate the performance of personalized recommendations. However, recent studies suggest that recommending a diverse list of products improves user satisfaction and is positively associated with customer retention rates. In this study, we propose to incorporate the product ratings from users’ online social relations into recommendation model to enhance the diversity of personalized recommendation list. The empirical results indicate that our proposed approach performs well in increasing the recommendation diversity while maintaining comparable level of accuracy. The findings offer practical implications for online businesses to leverage online social relations.

Text-Based Interactive Recommendation via Offline Reinforcement Learning

Interactive recommendation with natural-language feedback can provide richer user feedback and has demonstrated advantages over traditional recommender systems. However, the classical online paradigm involves iteratively collecting experience via interaction with users, which is expensive and risky. We consider an offline interactive recommendation to exploit arbitrary experience collected by multiple unknown policies. A direct application of policy learning with such fixed experience suffers from the distribution shift. To tackle this issue, we develop a behavior-agnostic off-policy correction framework to make offline interactive recommendation possible. Specifically, we leverage the conservative Q-function to perform off-policy evaluation, which enables learning effective policies from fixed datasets without further interactions. Empirical results on the simulator derived from real-world datasets demonstrate the effectiveness of our proposed offline training framework.

Improved Music Recommendation Algorithm for Deep Neural Network Based on Attention Mechanism

With the increasingly close combination of the Internet and people’s production and life, the total amount of global data and information also grows increasingly. In order to save users the time to find their favorite music among many music types, the music recommendation service arises at the historic moment and is widely concerned by scholars. Traditional music recommendation system based on the collaborative filtering algorithm has low recommendation accuracy, poor real-time performance, sparsity, system cold start, and so on. Moreover, the traditional music recommendation algorithm only simply uses user behavior characteristics and does not make good use of user history for listening to audio characteristics. In view of the above question, this section based on the attention mechanism of the deep neural network music recommendation algorithm, through the use of improved MFCC audio data preprocessing, the extracted audio combined with the user’s own portrait features, through the AIN RNN network recommended list, by learning user history listening to songs, improves the model-recommended accuracy.

The adaptation of the information system success model in recommender systems. The validation of the dual-coding theory

Introduction. Recommender systems are extensively deployed to provide online users with advisory services, and the design of recommender systems functional features has received substantial attention in academic studies. The social aspects of human–recommender systems interactions, however, have been less explored. Furthermore, measuring user satisfaction, though natural in a business environment, is often challenging for recommender systems research. Thus, it is assumed that the information system success model can be adapted to the system success measurement in recommender systems. Method. This study provides the first empirical test of an adaptation of the information system success model in the context of recommender systems. Additionally, the dual-coding theory argument was introduced to examine the model. Analysis. Based on the proposed model, two presentation types are compared. An experimental design is used and a questionnaire is developed to analysis. Results. The experiment's results show that the recommender system designed on the basis of the dual-coding theory is better than the traditional recommender system on all aspects of the facet. Conclusions. Perceived usefulness and the user satisfaction have a significant positive relationship both with continued intention to use and continued intention to provide.

A deep selective learning network for cross-domain recommendation

In the past two decades, recommendation system has been successfully applied to many e-commerce companies and is a ubiquitous part of today online entertainment. However, many single-domain recommendations suffer from the sparsity problems due to a lack of sufficient interactive data. In fact, user behaviors from different domains are usually relevant. Therefore, cross-domain ideas have been proposed to help alleviate the data sparsity issue in traditional single-domain recommender systems. Motivated by this, we design a deep selective learning network (DSLN) in this paper, for the scenario when domains have minimum or no common users DSLN firstly exploits reviews to profile the preference of users and characteristic of items. Then it selects useful user or item information from the auxiliary domain and transfers it to the target domain to solve the negative transfer problem, even though there may be no overlapping users or items between these two domains. In DSLN model, the selection of useful information is realized by the de-noising auto-encoder (DAE), which is shared between the auxiliary and target domains. By minimizing the reconstruction error of the DAE, on the one hand, only the useful information can be selected from the auxiliary domain; on the other hand, the latent representation of users and items in two domains can be learned. Our experiments on three cross-domain scenarios with different sparsity of Amazon review dataset show that, our proposed model gains 0.58% to 18.16% relative improvement compared to single-domain recommendation models, and from 1.05% to 19.4% relative improvement compared to cross-domain recommendation models.

Personalized recommendation system based on social tags in the era of Internet of Things

Abstract With the rapid development of the Internet, recommendation systems have received widespread attention as an effective way to solve information overload. Social tagging technology can both reflect users’ interests and describe the characteristics of the items themselves, making group recommendation thus becoming a recommendation technology in urgent demand nowadays. In traditional tag-based recommendation systems, the general processing method is to calculate the similarity and then rank the recommended items according to the similarity. Without considering the influence of continuous user behavior, in this article, we propose a personalized recommendation algorithm based on social tags by combining the ideas of Markov chain and collaborative filtering. This algorithm splits the three-dimensional relationship of &lt;user-tag-item&gt; into two two-dimensional relationships of &lt;user-tag&gt; and &lt;tag-item&gt;. The user’s interest degree to the tags is calculated by the Markov chain model, and then the items corresponding to them are matched by the recommended tag set. The influence between tags is used to model the satisfaction of items based on the correlation between the tags contained in the matched items, and collaborative filtering is used to complete the sparse values when calculating the interest and satisfaction between user–tags and user–items to improve the accuracy of recommendations. The experiments show that in the publicly available dataset, the personalized recommendation algorithm proposed in this article has significantly improved in accuracy and recall rate compared with the existing algorithms.

Digital Movie Recommendation Algorithm Based on Big Data Platform

How to associate films with users among various film data and help users get useful information is a big problem we face. The recommendation system aims to provide users with accurate project recommendations, which can effectively solve the problem of information explosion caused by a large amount of data. Traditional recommendation systems are widely used in movie shopping. Aiming at this problem, this paper designs and develops a collaborative filtering recommendation algorithm based on big data platform. Firstly, the depth is deeper than the traditional automatic coding network, and the new activation function is used to generate the depth feature vector. Secondly, the model can describe both linear and nonlinear features of movie data, which further improves the extraction ability of nonlinear features. Experimental results show that the proposed algorithm is effective and can bring better user experience and economic benefits to consumers.

Deep Reinforcement Factorization Machines: A Deep Reinforcement Learning Model with Random Exploration Strategy and High Deployment Efficiency

In recent years, the recommendation system and robot learning are undoubtedly the two most popular application fields, and the core algorithms supporting these two fields are deep learning based on perception and reinforcement learning based on exploration learning, respectively. How to combine these two fields to better improve the development of the whole machine learning field is the dream of numerous researchers. The Deep Reinforcement Network (DRN) model successfully embedded reinforcement learning into the recommendation system, which provided a good idea for subsequent researchers. However, the disadvantage is also obvious, that is, the DRN model is built for news recommendations, meaning that the DRN model is not transferable, which is also the defect of many current recommendation system models. Meanwhile, the agent learning method adopted by the DRN model is primitive and inefficient. Among many models and algorithms that have emerged in recent years, we use the newly proposed deployment efficiency to measure their comprehensive quality and found that few models focus on both efficiency and performance improvement. To fill the gap of model deployment efficiency neglected by many researchers and to create a model of reinforcement learning agents with stronger performance, we have been exploring and trying to complete research on the Gate Attentional Factorization Machines (GAFM) model. Finally, we successfully integrated the GAFM model and reinforcement learning. The Deep Reinforcement Factorization Machines (DRFM) model proposed in this paper is based on the combination of deep learning with high perception ability and reinforcement learning with high exploration ability, centered on improving the deployment efficiency and learning performance of the model. The GAFM model is modified and upgraded using multidisciplinary techniques, and a new model-based random exploration strategy is proposed to update and optimize the recommendation list efficiently. Through parallel contrast experiments on various datasets, it is proved that the DRFM model surpasses the traditional recommendation system model in all aspects. The DRFM model is far superior to other models in terms of performance and robustness, and also significantly improved in terms of deployment efficiency. At the same time, we conduct a comparative analysis with the latest deep reinforcement learning algorithm and prove the unique advantages of the DRFM model.

Differentially private recommender system with variational autoencoders

To provide precise recommendations, traditional recommender systems (RS) collect personal data, user preference and feedback, which are sensitive to each user if such information is maliciously used for extra analysis. In recent years, differential privacy (DP) has been widely applied in RS to provide privacy protection for sensitive information. Prior studies explored the combination of DP and RS, while neglecting the disparate effect on model accuracy of imbalanced subgroups as large user groups control the trained model, and DP can worsen the disparate effect of degrading the performance of recommender systems significantly. Besides, the number of uploaded contributions can differ among users for training a recommender system, so it is necessary to set the user-level privacy guarantee.In this paper, we make four contributions. First, we propose an efficient way of constructing datasets for training a recommender system based on prior theories. Second, we compute the user-level priors based on user metadata to optimize the VAE model. Besides, we add noise into the calculation process to protect user metadata. Third, we analyze and propose a tighter theoretical bound on gradient updates for DP Stochastic Gradient Descent (DPSGD). Finally, we exploit these theoretical results and propose a novel DP-VAE based recommender system. Extensive experimental results on multiple datasets show that our system can achieve high recommendation precision while maintaining a reasonable privacy guarantee.

Context-aware Multi-stakeholder Recommender Systems

Traditional recommender systems help users find the most relevant products or services to match their needs and preferences. However, they overlook the preferences of other sides of the market (aka stakeholders) involved in the system. In this paper, we propose to use contextual bandit algorithms in multi-stakeholder platforms where a multi-sided relevance function with adjusting weights is modeled to consider the preferences of all involved stakeholders. This algorithm sequentially recommends the items based on the contextual features of users along with the priority of the stakeholders and their relevance to the items.Our extensive experimental results on a dataset consisting of MovieLens (1m), IMDB (81k+), and a synthetic dataset show that our proposed approach outperforms the baseline methods and provides a good trade-off between the satisfaction of different stakeholders over time.

Deep Conversational Recommender Systems: Challenges and Opportunities

Unlike traditional recommender systems, the conversational recommender system (CRS) models a user’s preferences through interactive dialogue conversations. Recently, deep learning approaches have been applied to CRSs, producing fruitful results. We discuss the development of deep CRSs and future research directions.