CF-based Recommender Systems Research Articles

TrustDL: Use of trust-based dictionary learning to facilitate recommendation in social networks

Collaborative filtering (CF) is a widely applied method to perform recommendation tasks in a wide range of domains and applications. Dictionary learning (DL) models, which are highly important in CF-based recommender systems (RSs), are well represented by rating matrices. However, these methods alone do not resolve the cold start and data sparsity issues in RSs. We observed a significant improvement in rating results by adding trust information on the social network. For that purpose, we proposed a new dictionary learning technique based on trust information, called TrustDL, where the social network data were employed in the process of recommendation based on structural details on the trusted network. TrustDL sought to integrate the sources of information, including trust statements and ratings, into the recommendation model to mitigate both problems of cold start and data sparsity. It conducted dictionary learning and trust embedding simultaneously to predict unknown rating values. In this paper, the dictionary learning technique was integrated into rating learning, along with the trust consistency regularization term designed to offer a more accurate understanding of the feature representation. Moreover, partially identical trust embedding was developed, where users with similar rating sets could cluster together, and those with similar rating sets could be represented collaboratively. The proposed strategy appears significantly beneficial based on experiments conducted on four frequently used datasets: Epinions, Ciao, FilmTrust, and Flixster.

Autoencoder-Based Recommender System Exploiting Natural Noise Removal

Collaborative filtering (CF) is a widely used technique in recommender systems by automatically predicting the user’s latent interests based on many users’ historical rating data. To improve the performance of the CF-based recommender systems, users’ rating data should be pre-processed to avoid noise and enhance data reliability. Many researchers studied anomaly detection to remove malicious noise caused by shilling attacks, but anomalies can still exist in non-attacked real user data, which is called natural noise, as the ratings of users can be impacted by unpredictable factors such as other users’ ratings and anchoring bias. In this paper, we propose an autoencoder-based recommendation system for exploiting the ability of both anomaly detection and CF. The proposed system detects the natural noise in the rating data based on the reconstruction errors after training. By removing the detected natural noise, CF can predict the unrated ratings with noise-free data. Our experiments show that the proposed model showed better performance than the traditional method by reducing the error by up to 5% compared to the method that does not consider natural noise detection and reducing the error by up to 4% compared to the conventional rating classification based natural noise detection methods.

A Learning Automata-Based Approach to Improve the Scalability of Clustering-Based Recommender Systems

One of the common techniques to reduce the scalability problem in collaborative filtering (CF)-based recommender systems is the clustering technique, which accelerates finding the nearest neighbor users in the recommendation process. Different clustering algorithms lead to improved accuracy and diversity in recommender systems. It is challenging to develop recommender systems based on clustering with decreasing scalability and simultaneously increasing accuracy. This article proposes a new clustering-based recommender system that takes into account the theoretical properties of the Learning Automata (LA) technique. The presented clustering novelty lies in the fact that employing LA technique for the user clustering in a CF-based recommender system has not been set forth so far in a way that addresses the scalability while improving accuracy. In addition, a novel similarity metric is embedded in the proposed algorithm to measure the similarity value between users. This metric is developed as like/dislike (LD) which can significantly improve the accuracy by reducing the computational cost. Extensive simulations have been performed on real-world datasets such as MovieLens and FilmTrust, which confirm the effectiveness of the proposed algorithm. In this regard, the proposed algorithm has improved the precision between 5 and 16% on average compared to the existing state-of-the-art methods such as GA-GELS, NUSCCF, NNMF, and KL-KM.

Enhanced Collaborative Filtering for Personalized E-Government Recommendation

The problems with the information overload of e-government websites have been a big obstacle for users to make decisions. One promising approach to solve this problem is to deploy an intelligent recommendation system on e-government platforms. Collaborative filtering (CF) has shown its superiority by characterizing both items and users by the latent features inferred from the user–item interaction matrix. A fundamental challenge is to enhance the expression of the user or/and item embedding latent features from the implicit feedback. This problem negatively affected the performance of the recommendation system in e-government. In this paper, we firstly propose to learn positive items’ latent features by leveraging both the negative item information and the original embedding features. We present the negative items mixed collaborative filtering (NMCF) method to enhance the CF-based recommender system. Such mixing information is beneficial for extending the expressiveness of the latent features. Comprehensive experimentation on a real-world e-government dataset showed that our approach improved the performance significantly compared with the state-of-the-art baseline algorithms.

Unifying user similarity and social trust to generate powerful recommendations for smart cities using collaborating filtering-based recommender systems

Recommender systems can improve the quality of life in smart cities by presenting personalized services to the community. Such systems maintain a database of user profiles for producing recommendations for a specific user. The collaborative filtering (CF) approach used in these systems has become a benchmark approach for generating recommendations for interested users because it can provide “out of the box” solutions. These CF-based approaches first construct a user–item rating matrix and then exploit similarity methods. These approaches suffer from scalability, sparsity, and cold user conditions, which consequently result in the poor recommendation accuracy of these systems. To enhance the accuracy of recommender systems, social trust can play a vital role because people tend to interact with a system or respond positively to recommendations that originate from their social trustworthy friends. The proposed unified approach of this article uses explicit trust, implicit trust, and user preference similarity to create a unified rating profile for the target user to produce more powerful and accurate recommendations. The proposed unified approach also enhances the recommendation performance of CF-based recommender systems when only a limited set of ratings is available. Experiments are performed on three publicly available datasets which are FilmTrust, CiaoDVD, and Epinions. Comparison of obtained results is made with traditional similarity measures as well as up-to-date trust-based approaches. The results show that the proposed unified approach is superior to existing approaches in terms of both predictive and classification-based accuracy measures.

Modeling user rating preference behavior to improve the performance of the collaborative filtering based recommender systems

One of the main concerns for online shopping websites is to provide efficient and customized recommendations to a very large number of users based on their preferences. Collaborative filtering (CF) is the most famous type of recommender system method to provide personalized recommendations to users. CF generates recommendations by identifying clusters of similar users or items from the user-item rating matrix. This cluster of similar users or items is generally identified by using some similarity measurement method. Among numerous proposed similarity measure methods by researchers, the Pearson correlation coefficient (PCC) is a commonly used similarity measure method for CF-based recommender systems. The standard PCC suffers some inherent limitations and ignores user rating preference behavior (RPB). Typically, users have different RPB, where some users may give the same rating to various items without liking the items and some users may tend to give average rating albeit liking the items. Traditional similarity measure methods (including PCC) do not consider this rating pattern of users. In this article, we present a novel similarity measure method to consider user RPB while calculating similarity among users. The proposed similarity measure method state user RPB as a function of user average rating value, and variance or standard deviation. The user RPB is then combined with an improved model of standard PCC to form an improved similarity measure method for CF-based recommender systems. The proposed similarity measure is named as improved PCC weighted with RPB (IPWR). The qualitative and quantitative analysis of the IPWR similarity measure method is performed using five state-of-the-art datasets (i.e. Epinions, MovieLens-100K, MovieLens-1M, CiaoDVD, and MovieTweetings). The IPWR similarity measure method performs better than state-of-the-art similarity measure methods in terms of mean absolute error (MAE), root mean square error (RMSE), precision, recall, and F-measure.

A big-data oriented recommendation method based on multi-objective optimization

Due to its successful application in recommender systems, collaborative filtering (CF) has become a hot research topic in data mining. For traditional CF-based recommender systems, the accuracy of recommendation results can be guaranteed while the diversity will be lost. An ideal recommender system should be built with both accurate and diverse performance. Faced with accuracy–diversity dilemma, we propose a novel recommendation method based on MapReduce framework. In MapReduce framework, a block computational technique is used to shorten the operational time. And an improved collaborative filtering model is refined with a novel similarity computational process which considers many factors. By translating the procedure of generating personalized recommendation results into a multi-objective optimization problem, the multiple conflicts between accuracy and diversity are well handled. The experimental results demonstrate that our method outperforms other state-of-the-art methods.

Multi-criteria matrix localization and integration for personalized collaborative filtering in IoT environments

Collaborative filtering (CF)-based recommender systems can be used to deal with the complexity problem of users when they want to identify possible tasks on the fly and perform desired tasks by using various smart objects in Internet of Things (IoT) environments. However, in order to use CF-based recommender systems, users need to provide their feedbacks and there are usually more than one criterion considered when users choose an item. Although there have been studies of multi-criteria recommendations, existing approaches require multi-criteria ratings that are explicitly given by users. It is usually a burden for a user to provide more than one instance of feedback on an item; therefore, user feedback datasets are usually sparse when users are asked to provide multi-criteria ratings. Due to the sparsity of multi-criteria rating data, the similarity measurements used by the existing approaches may produce biased results, possibly leading to degradation of the recommendation accuracy. This problem becomes worse as the sparsity of a dataset increases. To alleviate the effects of the data-sparsity problem, and to take advantage of using multi-criteria ratings, we proposed a multi-criteria matrix localization and integration (MCMLI) approach for collaborative filtering in this paper. The main goal of MCMLI is to find cohesive user-item subgroups (CUISs) for each criterion from sparse data, and to predict users’ interests for each criterion in a more precise manner. The proposed approach is composed of three phases. At the first phase, a given user-item matrix is divided into a set of CUIS matrices, each of which is organized with correlated users and items for each criterion. MCMLI repeats this CUIS generation process until the generated subgroups cover all elements of the given user-item matrix. To generate prediction results for each criterion, MCMLI then predicts user ratings on new items for each CUIS and aggregates the prediction results to make recommendations to users. To enable personalized recommendations, during the aggregation process, each user’s preferences on multiple criteria are weighted differently according to the number of CUISs to which the user belongs. We demonstrate the effectiveness of our approach by conducting an experiment with real-world datasets from TripAdvisor and Yahoo! Movies. The experimental results show that MCMLI outperforms existing multi-criteria collaborative-filtering-based recommendation methods in terms of the recommendation accuracy. In addition, unlike the existing multi-criteria recommendation approaches, even when the sparsity level of a dataset increases, the recommendation accuracy of MCMLI does not decrease significantly.

Recommender systems using cluster-indexing collaborative filtering and social data analytics

As a result of the extensive variety of products available in e-commerce settings during the last decade, recommender systems have been highlighted as a means of mitigating the problem of information overload. Collaborative filtering (CF) is the most widely used algorithm to build such systems, and improving the predictive accuracy of CF-based recommender systems has been a major research challenge. This research aims to improve the prediction accuracy of CF by incorporating social network analysis (SNA) and clustering techniques. Our proposed model identifies the most influential people in an online social network by SNA and then conducts clustering analysis using these people as initial centroids (cluster centres). Finally, the model makes recommendations using cluster-indexing CF based on the clustering outcomes. In this step, our model adjusts the effect of neighbours in the same cluster as the target user to improve prediction accuracy by reflecting hidden information about his or her social community. The experimental results indicate that the proposed model outperforms other comparison models, including conventional CF, with statistical significance.

A semantic-enhanced trust based recommender system using ant colony optimization

Collaborative Filtering (CF) is the most popular recommendation technique that uses preferences of users in a community to make personal recommendations for other users. Despite its popularity and success, CF suffers from the data sparsity and cold-start problems. To alleviate these issues, in recent years, there has been an upsurge of interest in exploiting trust information to improve the performance of CF. In general, trust has a number of distinct properties such as asymmetry, transitivity, dynamicity and context-dependency. However, conventional trust-based CF systems do not address trust computation by considering all the properties of trust. Particularly, the context-dependency property has received less attention in the existing approaches. The consideration of all these properties leads to more accurate recommendations since the quality of the inferred is improved. In this paper, we propose a novel trust-based approach, called Semantic-enhanced Trust based Ant Recommender System (STARS), which satisfies all the properties mentioned above. Using ant colony optimization, the proposed system performs a depth- first search for the optimal trust paths in the trust network and selects the best neighbors of an active user to provide better recommendations. To consider the context-dependency property, trust inference in STARS depends on the semantic descriptions of items. Incorporation of both global and local trust in CF-based recommender systems in addition to the trust computation based on the semantic features of items allows STARS to alleviate the data sparsity, cold-start and "multiple-interests and multiple-content" problems of CF. Experimental results on real-world data sets show that STARS outperforms its counterparts in terms of prediction accuracy and recommendation quality and can overcome the above problems.

Improving Collaborative Recommendation via User-Item Subgroups

Collaborative filtering (CF) is out of question the most widely adopted and successful recommendation approach. A typical CF-based recommender system associates a user with a group of like-minded users based on their individual preferences over all the items, either explicit or implicit, and then recommends to the user some unobserved items enjoyed by the group. However, we find that two users with similar tastes on one item subset may have totally different tastes on another set. In other words, there exist many user-item subgroups each consisting of a subset of items and a group of like-minded users on these items. It is more reasonable to predict preferences through one user's correlated subgroups, but not the entire user-item matrix. In this paper, to find meaningful subgroups, we formulate a new Multiclass Co-Clustering (MCoC) model, which captures relations of user-to-item, user-to-user, and item-to-item simultaneously. Then, we combine traditional CF algorithms with subgroups for improving their top- $N$ recommendation performance. Our approach can be seen as a new extension of traditional clustering CF models. Systematic experiments on several real data sets have demonstrated the effectiveness of our proposed approach.

Opinion-enhanced collaborative filtering for recommender systems through sentiment analysis

The motivation of collaborative filtering (CF) comes from the idea that people often get the best recommendations from someone with similar tastes. With the growing popularity of opinion-rich resources such as online reviews, new opportunities arise as we can identify the preferences from user opinions. The main idea of our approach is to elicit user opinions from online reviews, and map such opinions into preferences that can be understood by CF-based recommender systems. We divide recommender systems into two types depending on the number of product category recommended: the multiple-category recommendation and the single-category recommendation. For the former, sentiment polarity in coarse-grained manner is identified while for the latter fine-grained sentiment analysis is conducted for each product aspect. If the evaluation frequency for an aspect by a user is greater than the average frequency by all users, it indicates that the user is more concerned with that aspect. If a user's rating for an aspect is lower than the average rating by all users, he or she is much pickier than others on that aspect. Through sentiment analysis, we then build an opinion-enhanced user preference model, where the higher the similarity between user opinions the more consistent preferences between users are. Experiment results show that the proposed CF algorithm outperforms baseline methods for product recommendation in terms of accuracy and recall.

An Adaptive Recommendation Method Based on Small-World Implicit Trust Network

Collaborative filtering (CF) is widely used in e-commerce recommender systems, which helps the online users to identify the right products to purchase. However, CF-based recommender systems suffer poor quality of recommendation due to the sparsity issue. To address this problem, in this paper we propose an adaptive recommendation method based on small-world implicit trust network. We first present a method to construct the small-world implicit trust network based on user clustering and implicit trust among users. Then we develop an adaptive recommendation algorithm by taking into account the topology of the constructed trust network, which generates recommendations using different strategies. To demonstrate the effectiveness of the proposed method, we conduct experiments on the MovieLens dataset and compare our method with others. Experimental results show that the proposed method can significantly improve the quality of recommendation.

Exploiting two-faceted web of trust for enhanced-quality recommendations

Traditional collaborative filtering (CF) based recommender systems on the basis of user similarity often suffer from low accuracy because of the difficulty in finding similar users. Incorporating trust network into CF-based recommender system is an attractive approach to resolve the neighbor selection problem. Most existing trust-based CF methods assume that underlying relationships (whether inferred or pre-existing) can be described and reasoned in a web of trust. However, in online sharing communities or e-commerce sites, a web of trust is not always available and is typically sparse. The limited and sparse web of trust strongly affects the quality of recommendation. In this paper, we propose a novel method that establishes and exploits a two-faceted web of trust on the basis of users’ personal activities and relationship networks in online sharing communities or e-commerce sites, to provide enhanced-quality recommendations. The developed web of trust consists of interest similarity graphs and directed trust graphs and mitigates the sparsity of web of trust. Moreover, the proposed method captures the temporal nature of trust and interest by dynamically updating the two-faceted web of trust. Furthermore, this method adapts to the differences in user rating scales by using a modified Resnick’s prediction formula. As enabled by the Pareto principle and graph theory, new users highly benefit from the aggregated global interest similarity (popularity) in interest similarity graph and the global trust (reputation) in the directed trust graph. The experiments on two datasets with different sparsity levels (i.e., Jester and MovieLens datasets) show that the proposed approach can significantly improve the predictive accuracy and decision-support accuracy of the trust-based CF recommender system.

Recommender Systems as a Mobile Marketing Service

Recommender systems (RSs) have become a familiar artifact in cyberspace as a vehicle for increasing revenues while deepening customer loyalty and satisfaction. Typically RS are developed in house by companies with a large product line and customer base. However, the structure of RS is often straightforward, and effective systems can be developed at relatively low cost, and thus offered as a marketing service. We discuss the anatomy of a specific recommender system designed for a telecommunication carrier which uses predictive analytics in the form of collaborative filtering techniques to recommend products to users. Collaborative filtering (CF) is based upon the premise that users who have purchased a particular product will have similar preferences to other users who also purchased the product. We discuss and compare three versions of a CF-based recommender system based upon customer purchase history, customer browsing history, and user segments respectively. The results in terms of increased sales suggest that RS offer substantial value as a mobile marketing service. We suggest ways in which RS can be generalized into RS generators for more rapid development and deployment. We then invert the RS perspective from product-centric to user-centric and suggest how this would work in customer targeting for mobile advertising campaigns. We conclude that RS in the largest sense are heavily model-feedback in nature and require increasingly sophisticated automated modeling and predictive analytics capabilities layered on a scalable big data infrastructure.

Semantic clustering-based attack detection model on CF-based recommender systems

Collaborative recommender systems have been widely used in E-commerce environment.Because this recommendation technology is very sensitive to user's profile,an attacker can affect the prediction by injecting a lot of biased users' profiles.Therefore,the author proposed a semantic clustering-based attack detection model on CF-based recommender systems,which mined the potential interest combination by analyzing the semantics of items in the transaction database.The proposed model judged the truth of a user's profile by detecting the randomness in a user's data.Extensive experiments demonstrate that the proposed model can effectively detect the "profile injection" attacks in CF-based recommender system,which can significantly improve the robustness and reliability of the whole system.

A scalable P2P recommender system based on distributed collaborative filtering

Collaborative Filtering (CF) technique has been proved to be one of the most successful techniques in recommender systems in recent years. However, most existing CF based recommender systems worked in a centralized way and suffered from its shortage in scalability as their calculation complexity increased quickly both in time and space when the record in user database increases. In this article, we first propose a distributed CF algorithm called PipeCF together with two novel approaches: significance refinement and unanimous amplification, to further improve the scalability and prediction accuracy. We then show how to implement this algorithm on a Peer-to-Peer (P2P) structure through distributed hash table method, which is the most popular and efficient P2P routing algorithm, to construct a scalable distributed recommender system. The experimental data show that the distributed CF-based recommender system has much better scalability than traditional centralized ones with comparable prediction efficiency and accuracy. q 2004 Elsevier Ltd. All rights reserved.

Probabilistic memory-based collaborative filtering

Memory-based collaborative filtering (CF) has been studied extensively in the literature and has proven to be successful in various types of personalized recommender systems. In this paper, we develop a probabilistic framework for memory-based CF (PMCF). While this framework has clear links with classical memory-based CF, it allows us to find principled solutions to known problems of CF-based recommender systems. In particular, we show that a probabilistic active learning method can be used to actively query the user, thereby solving the "new user problem." Furthermore, the probabilistic framework allows us to reduce the computational cost of memory-based CF by working on a carefully selected subset of user profiles, while retaining high accuracy. We report experimental results based on two real-world data sets, which demonstrate that our proposed PMCF framework allows an accurate and efficient prediction of user preferences.

Application of Web usage mining and product taxonomy to collaborative recommendations in e-commerce

The rapid growth of e-commerce has caused product overload where customers on the Web are no longer able to effectively choose the products they are exposed to. To overcome the product overload of online shoppers, a variety of recommendation methods have been developed. Collaborative filtering (CF) is the most successful recommendation method, but its widespread use has exposed some well-known limitations, such as sparsity and scalability, which can lead to poor recommendations. This paper proposes a recommendation methodology based on Web usage mining, and product taxonomy to enhance the recommendation quality and the system performance of current CF-based recommender systems. Web usage mining populates the rating database by tracking customers' shopping behaviors on the Web, thereby leading to better quality recommendations. The product taxonomy is used to improve the performance of searching for nearest neighbors through dimensionality reduction of the rating database. Several experiments on real e-commerce data show that the proposed methodology provides higher quality recommendations and better performance than other CF methodologies.