Trust-aware Recommender Systems Research Articles

Suspiciousness and Fast and Slow Thinking Impact on Trust in Recommender Systems

Abstract By offering consumers more proactive and individualized information services, recommender systems have proven to be a significant answer to the problem of information overload. And collaborative filtering approaches have proven to be an important component of many such recommender systems, allowing for the development of high-quality recommendations by harnessing the preferences of communities of similar users. In this study, we argue that individual personality traits may play an essential role in ensuring general end-user confidence in recommender systems. Individual elements to be examined include social suspiciousness as a personality feature and Daniel Kahneman’s fast and slow thinking styles, which distinguish between two modes of thought: System 1 is rapid, instinctual, and emotive, while System 2 is slower, more deliberate, and more rational. Due to a lack of explanation or inaccurate recommendation results, users may not trust current recommender systems, which poses a significant challenge for those systems. As such, having a trustworthy recommender system is essential. Three different trust-aware recommender system types are analyzed systematically in this study: robust systems that filter misleading information such as spam and fake news; social aware conscious systems that benefit from users’ social connections; and explainable systems that offer justifications for recommended products. On a sample size of 487 Romanian respondents, in an online survey, we have analyzed the impact of individual factors on the trustworthiness of the three types of recommender systems. Results show that fast thinking fully mediates the relationship between social suspiciousness and trust in all three types of recommender systems.

A survey for trust-aware recommender systems: A deep learning perspective

A significant remaining challenge for existing recommender systems is that users may not trust recommender systems for either inaccurate recommendation or lack of explanation. Thus, it becomes critical to embrace a trustworthy recommender system. This survey provides a systematic summary of three categories of trust issues in recommender systems: social-aware recommender systems, which leverage users’ social trust relationships; robust recommender systems, which filter untruthful information, noises and enhance attack resistance; and explainable recommender systems, which provide explanations of the recommended items. We focus on the work based on deep learning techniques, which is an emerging area in the recommendation research.

A reliable deep representation learning to improve trust-aware recommendation systems

Deep neural networks have been extensively employed in many applications such as natural language processing and computer vision. They have attracted a lot of attention in designing recommender systems due to their recent significant achievements. These models are mainly used in recommender systems to generate the users’ latent features according to the input data such as user-item rating and trust matrixes. However, such input data are significantly sparse leading to reduce the effectiveness of deep learning models in generating the latent features. In addition, the existing deep learning based recommendation models ignore the reliability measure in their recommendation process. To address these challenges, in this paper, a trust-aware recommendation method based on deep sparse autoencoder is proposed. Specifically, an effective probabilistic model is proposed to determine how many ratings are required for each user to produce an accurate prediction. Moreover, for enhancing the rating profiles of users with insufficient ratings, an efficient mechanism is developed based on the utilization of implicit ratings. In this mechanism, the implicit ratings are selected based on a reliability measure to guarantee their effectiveness. Then, the reliable enhanced rating profiles and trust statements are considered as the input data of deep sparse autoencoder to generate the users’ latent features. Finally, a similarity function that utilizes the generated latent features is developed to obtain similarity values between users and make recommendations. Massive experiments are carried out on two datasets to assess the effectiveness of the proposed method in comparison to the-state-of-the-art models. The results of experiments demonstrate that the proposed method can significantly beat other models in producing more accurate recommendations.

Trust and Distrust based Cross-domain Recommender System

ABSTRACT A recommender system (RS) provides assistance for users to filter out items of their interest in the presence of millions of available items. The reason is to find out the likewise user with the assumption that if users have shared similar interest in the past then they may share the same in future. Collaborative filtering (CF) is the widely used recommendation algorithm due to its ease of use but suffers with the problems of sparsity and cold start problem. In this paper, we propose a trust and distrust-based cross domain context aware recommender system in the multi-agent environment which tries to reduce the problem of data sparsity in collaborative-filtering recommender system and improves coverage. Cross Domain Recommender System (CDRS) utilizes data from multiple domains to reduce the problem of sparsity. Moreover, the combination of trust and distrust in recommendation help to improve trustworthiness of generated recommendation. Distrust provides higher accuracy in recommendation by incorporating knowledge about the malicious users. Prototype of the system is developed using JADE and Java technology for the tourism domain consisting of restaurant, hotel, travel places and shopping places as sub-domains. The performance of the proposed trust and distrust-based cross domain recommender system is compared with the traditional approach of recommendation along with the cross domain approach and trust-based cross-domain approach in terms of accuracy and coverage. The results show that the proposed system outperforms in terms of both accuracy and coverage.

Context-aware recommender system using trust network

Context-Aware Recommender Systems (CARS) improve traditional Recommender Systems (RS) in a wide array of domains and applications. However, CARS suffer from several inherent issues such as data sparsity and cold start. Incorporating trust into recommender systems can handle these issues. Trust-aware recommender systems use information from social networks such as trust statements, which prove another valuable information source. This paper exploits the advantages of these two systems by incorporating both trust and context information. We propose a hybrid approach: Trust based Context aware Post Filtering Approach that uses trust statements as a rich information with context compensation method of contextual post-filtering approach. Our approach utilizes the relative average difference among the context on output of trust aware collaborative filtering by incorporating explicit and implicit trust information. We also use a confidence concept to remove non-confident users from the trust network before generating prediction. The performed experiments show that the proposed approach improves upon the standard RS and outperforms recommendation approaches on real world dataset.

The Role of Trust to Enhance the Recommendation System Based on Social Network

Recommendation systems or recommender system (RSs) is one of the hottest topics nowadays, which is widely utilized to predict an item to the end-user based on his/her preferences primary. Recommendation systems applied in many areas mainly in commercial applications. This work aims to collect evidence of utilizing social network information between users to enhance the quality of traditional recommendation system. It provides an overview of traditional and modern approaches used by RSs such as collaborative filter (CF) approach, content-based (CB) approach, and hybrid filter approach. CF is one of the most famous traditional approaches in RSs, which is facing many limitations due to the lack of information available during a performance such as Cold start, Sparsity and Shilling attack. Additionally, this content focused on the role of incorporating a trust relationship from the social network to enhance the weaknesses of CF and achieve better quality in the recommendation process. Trust-aware Recommendation Systems (TaRSs) is a modern approach proposed to overcome the limitations of CF recommendation system in a social network. The trust relationship between users can boost and enhance CF limitations. Many researchers are focusing on trust in the recommendation system but fewer works are highlighting the role of trust in the recommendation system. In the end, limitations, and open issues of the current picture of the recommendation system come across.

A novel community-based trust aware recommender systems for big data cloud service networks

Internet of Things (IoT) has enhanced the appearance of the information and communications technology (ICT) infrastructure of smart sustainable cities that emerge as a growth of the urban development. The potential growth of environment sustainabilities is coming into the existence of the insightful visions of ICT that include big data analytics as a computing prototype. It may draw a penetrative path to mitigate the environmental effects that utilize the natural resource to manage the intelligent infrastructure and services. Moreover, the sustainable development process plays a crucial role to catalyze IoT and big-data application that connects the economic growth with the quality of smart cities. The IoT-enabled big-data applications choose a suitable thematic analysis to direct the use of recommendation systems. It supports users in the irresistible task of examination through large quantities of data in order to select appropriate information or items. Traditional recommender systems rely on similar neighbors irrespective of their preferences/choices when computing the prediction and assuming the users to be autonomous and indistinguishable to disregard the social events between the social users which are highly not reliable. Based on this intuition in this framework, the architecture of community-based trust aware recommender systems (CB-TARS) is proposed. Users’ preferences are expressed by incorporating trusted neighbors within the community of the target user are merged in order to find similar preferences. In addition, the worth of merged ratings is measured by the confidence considering the number of ratings inside the community and the percentage of clashes between negative and positive views. Further, rating confidence is incorporated into the computation of user similarity. The prediction for an unrated item is computed by aggregating the ratings of similar users within the community. Experimental results on real-world data set validate that the proposed method surpasses other complements in terms of accuracy.

Trust information network in social Internet of things using trust-aware recommender systems

Social Internet of things is one of the most up-to-date research issues in the applications of Internet of things technologies. In social Internet of things, accuracy and reliability are standard features to discerning decisions. We assume that decision support systems based on social Internet of things could leverage research from recommender systems to achieve more stable performance. Therefore, we propose a trust-aware recommender systems suitable for social Internet of things. Trust-aware recommender systems adapt the concept of social networking service and utilize social interaction information. Trust information not only improves recommender systems from opinion spam problems but also more accurately predicts users’ preferences. We confirm that the performance of a recommender system becomes more improved when implicit trust is able to satisfy the properties of trust in the social Internet of things environment. The structure and amount of social link information are context-sensitive, so applying the concept of trust into social Internet of things environments requires a method to optimize implicit and explicit trust with minimal social link information. Our proposed method configures an asymmetric implicit trust network utilizing user–item rating matrix and transforms trust propagation metrics for a directional and weighted trust network. Through experiments, we confirm that the proposed methods enable higher accuracy and wider coverage compared to the existing recommendation methods.

Extra: Expertise-Boosted Model for Trust-Based Recommendation System Based on Supervised Random Walk

The quality of recommendations based on any class of recommender systems may become poor if no or low quality data has been provided by users.This is a situation known as Cold Start problem, which typically happens when a new user registers to the system and no preference data is available for that user. Trust-Aware Recommendation Systems can be considered as a a solution for the cold start problem. In these systems, the trust between users plays an import role for making recommendations. However, most of the Trust-Aware RSs consider trust as a context independent phenomenon which means if user a trusts user b to the degree k then user a trusts user b to the degree k in all the concepts. However, in reality, trust is context dependent and user a can trust user b in context X but not in Y . Moreover, most of the trust-aware RSs do not consider an expertise concept for users and all the users are considered as same in the recommendation process. In this paper we proposed a novel approach for detecting expert users just based on their ratings (unlike previous systems which consider the separate profile and extra information for each user to nd an expert). In this model a supervised random walk is exploited to search the trust network for finding experts. Empirical experiments on the Epinions dataset shows that EXTRA can outperform previous models in terms of accuracy and coverage.

A Factored Similarity Model with Trust and Social Influence for Top-N Recommendation

Many trust-aware recommendation systems have emerged to overcome the problem of data sparsity, which bottlenecks the performance of traditional Collaborative Filtering (CF) recommendation algorithms. However, these systems most rely on the binary social network information, failing to consider the variety of trust values between users. To make up for the defect, this paper designs a novel Top-N recommendation model based on trust and social influence, in which the most influential users are determined by the Improved Structural Holes (ISH) method. Specifically, the features in Matrix Factorization (MF) were configured by deep learning rather than random initialization, which has a negative impact on prediction of item rating. In addition, a trust measurement model was created to quantify the strength of implicit trust. The experimental result shows that our approach can solve the adverse impacts of data sparsity and enhance the recommendation accuracy.

A significance-based trust-aware recommendation approach

Trust-aware recommender systems have been widely used in recent years to improve the performance of traditional collaborative filtering systems. A common assumption of existing trust models is that all items have the same importance for all users. However, it is reasonable to expect that some items are more significant than others in making recommendations. Furthermore, the significance of an item is not the same for all users but varies depending on many factors such as the demographic characteristics of users. For example, an item that is important to women may not be important to men. Also, the significance of an item for an individual user is not static and can change throughout the life cycle (from childhood to old age). Thus, items that are currently important to a user may become less important in the future. In this paper, we propose a Significance-Based Trust-Aware Recommendation (SBTAR) approach, which uses a new trust measure based on the concept of item significance. The significance of an item for a user is measured with respect to the demographic context of the user. Thus, SBTAR can adapt to dynamic changes in user preferences. To model demographic context, SBTAR uses Shuffled Frog Leaping Algorithm (SFLA), which is a meta-heuristic optimization technique based on the social behavior of frogs. SFLA has the advantages of simplicity, fast convergence, strong global search ability and easy implementation. Experimental results show that the proposed approach is more effective and efficient than several state-of-the-art recommendation approaches.

Learning adaptive trust strength with user roles of truster and trustee for trust-aware recommender systems

There are two key characteristics of users in trust relationships that have been well studied: (1) users trust their friends with different trust strengths and (2) users play multiple roles of trusters and trustees in trust relationships. However, few studies have considered both of these factors. Indeed, it is quite common for someone to respond to his/her friend that they trusted him/her, which indicates that there exist two kinds of information between each pair of users: the trust influence of trustee on truster and the feedback influence of truster on trustee. Considering this problem, we propose a novel adaptive method to learn the trust influence between users with multiple roles of truster and trustee for recommendation. First, we propose to introduce the concept of latent trust strength to learn adaptive role-based trust strength with limited values for each trust relationship between users. Second, because there is only one training example to learn each parameter of latent trust strength, we further propose two regularization methods by building relations between latent trust strength and user preferences to guide the training process of latent trust strength. After that, we develop a new recommendation method, RoleTS, by integrating the role-based trust strength into a previous recommendation model, TrustSVD, which considers both explicit and implicit information of trust and ratings. We also conduct a series of experiments to study the performance of the proposed method. Experimental results on two public real datasets demonstrate that the proposed method performs better than several state-of-the-art algorithms.

A Dynamic Local–Global Trust-aware Recommendation approach

Abstract Trust-aware recommender systems have attracted much attention in recent years because of the popularity of social networks. Some researchers have proposed local trust models to measure trust between two users based on their interactions. However, in most cases, there are few (if any) direct interactions between two users. In such cases, it is useful to associate a user with a global reputation value that reflects the experience of the whole community with that user. In this paper, we propose a Dynamic Local–Global Trust-aware Recommendation (DLGTR) approach, which uses a new hybrid model of local and global trust. In this hybrid model, the relative importance of local trust versus global trust is dynamically determined based on the reliability of information from the two trust models. Moreover, DLGTR uses an incremental method to efficiently update a trust network as new data become available. This allows DLGTR to constantly improve the recommendation quality without loss of efficiency. Experimental results demonstrate that the proposed approach outperforms the state-of-the-art recommendation approaches, especially in terms of prediction accuracy and computational time.

Low-Complexity Decentralized Recommendation System With Similarity Constraints

In recommendation systems (RSs), nowadays, not only the traditional user-item rating matrices but also more additional information like contents, contexts, trust friends and other auxiliary information are available to enhance the performance of RS, leading to content-aware, context-aware, trust-aware RS, etc. Thus, it provides much potential to take into consideration the additional information in RS. Hence, we focus on a general low-rank matrix factorization (LRMF) model with similarity constraints and propose a decentralized algorithm based on alternating direction method of multipliers (ADMM) to relieve the computation burden in each server while preserving privacy. What’s more, we utilize low-complexity skills in numerical analysis to reduce the computational complexity, based on the exploitation of the special form of the problem. Finally, simulations are performed to validate the effectiveness of our algorithms.

An Evaluation Framework of Trust Aware Recommender System

To date, there exists a variety of prediction approaches have been used in recommender systems. Among the widely known approaches are Content Based Filtering (CBF) and Collaborative Filtering (CF). Based on literatures, CF with users rating element has been widely used but the approach faced two common problems namely cold start and sparsity. As an alternative, Trust Aware Recommender Systems (TARS) for the CF based users rating has been introduced. The research progress on TARS improvement is found to be rapidly progressing but lacking in the algorithm evaluation has been started to appear. Many researchers that introduced their new TARS approach provides different evaluation of users’ views for the TARS performances. As a result, the performances of different TARS from different publications are not comparable and difficult to be analyzed. Therefore, this paper is written with objective to provide common group of the users’ views based on trusted users in TARS. Then, this paper demonstrates a comparison study between different TARS techniques with the identified common groups by means of the accuracy error, rating and users coverage. The results therefore provide a relative comparison between different TARS.

An effective social recommendation method based on user reputation model and rating profile enhancement

Trust-aware recommender systems are advanced approaches which have been developed based on social information to provide relevant suggestions to users. These systems can alleviate cold start and data sparsity problems in recommendation methods through trust relations. However, the lack of sufficient trust information can reduce the efficiency of these methods. Moreover, diversity and novelty are important measures for providing more attractive suggestions to users. In this article, a reputation-based approach is proposed to improve trust-aware recommender systems by enhancing rating profiles of the users who have insufficient ratings and trust information. In particular, we use a user reliability measure to determine the effectiveness of the rating profiles and trust networks of users in predicting unseen items. Then, a novel user reputation model is introduced based on the combination of the rating profiles and trust networks. The main idea of the proposed method is to enhance the rating profiles of the users who have low user reliability measure by adding a number of virtual ratings. To this end, the proposed user reputation model is used to predict the virtual ratings. In addition, the diversity, novelty and reliability measures of items are considered in the proposed rating profile enhancement mechanism. Therefore, the proposed method can improve the recommender systems about the cold start and data sparsity problems and also the diversity, novelty and reliability measures. Experimental results based on three real-world datasets show that the proposed method achieves higher performance than other recommendation methods.

Trust aware recommender system using swarm intelligence

Due to limitations and challenges faced by traditional collaborative filtering-based recommender systems, researchers have been shifting their attention towards using trust information among users while generating recommendations. It is observed that one trust metric may work better for some user and fails to do so in the case of another user. This paper proposes to favor that metric which provides high-quality recommendations for a particular user. For this purpose, weights have been assigned to various trust metrics for each pair of users and optimized iteratively to generate more accurate and personalized recommendations. We have used swarm intelligent techniques namely Bat algorithm and Particle Swarm Optimization for the same. The performance of the approach proposed in this work is evaluated using MovieLens, Epinions, CiaoDVD, and Filmtrust data sets and compared to earlier works generating recommendations using an individual metric. The results indicate that the MAE while using a combination of weighted trust is 0.59 (with PSO) and 0.55 (with Bat), which is much better compared to using a single metric. Bat also generates better recommendations with an accuracy of 85.45% than PSO (81.85%). Also, the MAE found using Bat was 3.84% better when utilizing denser datasets (MovieLens and FilmTrust) as compared to sparse datasets (CiaoDVD and Epinions).

A unified framework of trust prediction based on message passing

Trust prediction has been a vital part in trust-aware recommender systems. There are two approaches to predict trust between users: an explicit-information approach and an implicit-information approach. The explicit-information approach uses the explicit trust information observed from a trustor-trustee pair, whereas the implicit-information approach uses the implicit-information inferred from interaction features between a candidate trustor-trustee pair. The existing researches on trust prediction have been mainly focusing on one kind of information rather than combining the two kinds collectively. The task of trust prediction, however, could be improved when all kinds of information available from user pairs are integrated. Thus, we develop a new framework, including trust propagation mechanisms based on trust-message passing which takes advantage of the two kinds of information. In the developed framework, we build various probability-based trust prediction models according to the way of integrating the trust propagation mechanisms. Using real-world data, we empirically determine a weight vector for the integration of the trust propagation mechanisms and select the best trust prediction model. We demonstrate the benefit of the proposed trust propagation mechanisms by showing that the proposed method, ITD, outperforms existing methods, ABIT_L and MoleTrust, by 12.5 and 29.1%, respectively, in accuracy.

Trust aware recommender system with distrust in different views of trusted users

A recommender system aims to provide users with personalized online product or service recommendations to handle the online information overload problem that keep rapidly increasing. The main problems in the CF recommender system are sparsity and cold start. In order to resolve the problems, one of the current researches has been directed to the CF with a trust aware mechanism that includes trust as additional information in order to predict the rating for sparse data. This paper provides a review of the existing recommender system implementing the CF and trust aware. Furthermore, based on an empirical experiment, the performances of two recommender system approaches with trust aware and distrust in different views of trusted users are also reported in this paper. The results have shown that the different views have an effect on the accuracy and rating coverage of the two algorithms.

EXTRA: EXpertise-Boosted Model for Trust-Based Recommendation System Based on Supervised Random Walk

The quality of recommendations based on any class of recommender systems may become poor if no or low quality data has been provided by users. This is a situation known as it Cold Start problem, which typically happens when a new user registers to the system and no preference data is available for that user. Trust-Aware Recommendation Systems can be considered as a solution for the cold start problem. In these systems, the trust between users plays an import role for making recommendations. However, most of the Trust-Aware RSs consider trust as a context independent phenomenon which means if user a trusts user b to the degree k then user a trusts user b to the degree k in all the concepts. However, in reality, trust is context dependent and user a can trust user b in context X but not in Y. Moreover, most of the trust-aware RSs do not consider an expertise concept for users and all the users are considered as same in the recommendation process. In this paper we proposed a novel approach for detecting expert users just based on their ratings (unlike previous systems which consider the separate profile and extra information for each user to find an expert). In this model a supervised random walk is exploited to search the trust network for finding experts. Empirical experiments on the Epinions dataset shows that EXTRA can outperform previous models in terms of accuracy and coverage.