融合用户信任模型的协同过滤推荐算法

协同过滤推荐是电子商务系统中最为重要的技术之一.随着电子商务系统中用户数目和商品数目的增加,用户-项目评分数据稀疏性问题日益显著.传统的相似度度量方法是基于用户共同评分项目计算的,而过于稀疏的评分使得不能准确预测用户偏好,导致推荐质量急剧下降.针对上述问题,本文考虑用户评分相似性和用户之间信任关系对推荐结果的影响,利用层次分析法实现用户信任模型的构建,提出一种融合用户信任模型的协同过滤推荐算法.实验结果表明： 该算法能够有效反映用户认知变化,缓解评分数据稀疏性对协同过滤推荐算法的影响,提高推荐结果的准确度.     

用户多兴趣下基于信任的协同过滤算法研究

协同过滤技术是目前电子商务推荐系统最为有效的信息过滤技术之一.最近的研究尝试在推荐过程中引入信任模型来提高推荐的准确性和抵御"托"攻击.但在用户多兴趣的情况下,属于不同主题的项目需要不同的可信赖人员来推荐,传统的概貌级信任模型已不再适用.本文提出主题级信任计算模型以及基于主题级信任的协同过滤算法.一系列的实验结果表明,该算法在不牺牲鲁棒性的同时,有效地提高了推荐的准确性.     

一种融合项目特征和移动用户信任关系的推荐算法

协同过滤推荐系统中普遍存在评分数据稀疏问题.传统的协同过滤推荐系统中的余弦、Pearson 等方法都是基于共同评分项目来计算用户间的相似度;而在稀疏的评分数据中,用户间共同评分的项目所占比重较小,不能准确地找到偏好相似的用户,从而影响协同过滤推荐的准确度.为了改变基于共同评分项目的用户相似度计算,使用推土机距离（earth mover's distance,简称EMD）实现跨项目的移动用户相似度计算,提出了一种融合项目特征和移动用户信任关系的协同过滤推荐算法.实验结果表明：与余弦、Pearson 方法相比,融合项目特征的用户相似度计算方法能够缓解评分数据稀疏对协同过滤算法的影响.所提出的推荐算法能够提高移动推荐的准确度.     

Collaborative filtering and deep learning based recommendation system for cold start items

Recommender system is a specific type of intelligent systems, which exploits historical user ratings on items and/or auxiliary information to make recommendations on items to the users. It plays a critical role in a wide range of online shopping, e-commercial services and social networking applications. Collaborative filtering (CF) is the most popular approaches used for recommender systems, but it suffers from complete cold start (CCS) problem where no rating record are available and incomplete cold start (ICS) problem where only a small number of rating records are available for some new items or users in the system. In this paper, we propose two recommendation models to solve the CCS and ICS problems for new items, which are based on a framework of tightly coupled CF approach and deep learning neural network. A specific deep neural network SADE is used to extract the content features of the items. The state of the art CF model, timeSVD++, which models and utilizes temporal dynamics of user preferences and item features, is modified to take the content features into prediction of ratings for cold start items. Extensive experiments on a large Netflix rating dataset of movies are performed, which show that our proposed recommendation models largely outperform the baseline models for rating prediction of cold start items. The two proposed recommendation models are also evaluated and compared on ICS items, and a flexible scheme of model retraining and switching is proposed to deal with the transition of items from cold start to non-cold start status. The experiment results on Netflix movie recommendation show the tight coupling of CF approach and deep learning neural network is feasible and very effective for cold start item recommendation. The design is general and can be applied to many other recommender systems for online shopping and social networking applications. The solution of cold start item problem can largely improve user experience and trust of recommender systems, and effectively promote cold start items.     

Preventing Recommendation Attack in Trust-Based Recommender Systems

Despite its success, similarity-based collaborative filtering suffers from some limitations, such as scalability, sparsity and recommendation attack. Prior work has shown incorporating trust mechanism into traditional collaborative filtering recommender systems can improve these limitations. We argue that trust-based recommender systems are facing novel recommendation attack which is different from the profile injection attacks in traditional recommender system. To the best of our knowledge, there has not any prior study on recommendation attack in a trust-based recommender system. We analyze the attack problem, and find that “victim” nodes play a significant role in the attack. Furthermore, we propose a data provenance method to trace malicious users and identify the “victim” nodes as distrust users of recommender system. Feasibility study of the defend method is done with the dataset crawled from Epinions website.     

Applications of wavelet data reduction in a recommender system

Memory-based collaborative filtering (CF) recommender systems have emerged as an effective technique for information filtering. CF recommenders are being widely adopted for e-commerce applications to assist users in finding and selecting items of interest. As a result, the scalability of CF recommenders presents a significant challenge; one that is particularly resilient because the volume of data these systems utilize will continue to increase over time. This paper examines the impact of discrete wavelet transformation (DWT) as an approach to enhance the scalability of memory-based collaborative filtering recommender systems. In particular, a wavelet transformation methodology is proposed and applied to both synthetic and real-world recommender ratings. For experimental purposes, the DWT methodology’s effect on predictive accuracy and calculation speed is evaluated to compare recommendation quality and performance.     

一种检测兴趣漂移的图结构推荐系统

协同过滤是构造推荐系统最有效的方法之一.其中,基于图结构推荐方法成为近来协同过滤的研究热点.基于图结构的方法视用户和项为图的结点,并利用图理论去计算用户和项之间的相似度.尽管人们对图结构推荐系统开展了很多的研究和应用,然而这些研究都认为用户的兴趣是保持不变的,所以不能够根据用户兴趣的相关变化做出合理推荐.本文提出一种新的可以检测用户兴趣漂移的图结构推荐系统.首先,设计了一个新的兴趣漂移检测方法,它可以有效地检测出用户兴趣在何时发生了哪种变化.其次,根据用户的兴趣序列,对评分项进行加权并构造用户特征向量.最后,整合二部投影与随机游走进行项推荐.在标准数据集MovieLens上的测试表明算法优于两个图结构推荐方法和一个评分时间加权的协同过滤方法.     

Typicality‐based collaborative filtering for book recommendation

Nowadays, personalized recommender system placed an important role to predict the customer needs, interest about particular product in various application domains, which is identified according to the product ratings. During this process, collaborative filtering (CF) has been utilized because it is one of familiar techniques in recommender systems. The conventional CF methods analyse historical interactions of user‐item pairs based on known ratings and then use these interactions to produce recommendations. The major challenge in CF is that it needs to calculate the similarity of each pair of users or items by observing the ratings of users on same item, whereas the typicality‐based CF determines the neighbours from user groups based on their typicality degree. Typicality‐based CF can predict the ratings of users with improved accuracy. However, to eliminate the cold start problem in the proposed recommender system, the demographic filtering method has been employed in addition to the typicality‐based CF. A weighted average scheme has been applied on the combined recommendation results of both typicality‐based CF and demographic‐based CF to produce the best recommendation result for the user. Thereby, the proposed system has been able to achieve a coverage ratio of more than 95%, which indicates that the system is able to provide better recommendation for the user from the available lot of products.     

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.     

基于差分隐私和SVD++的协同过滤算法

协同过滤技术在推荐系统的实现中具有广泛的应用,协同过滤以用户对商品项目的评价分数为依据,而这些评价有可能反映用户某些不欲为人知的喜好特点,因此,对具备隐私保护能力的协同过滤模型的研究引起了普遍的关注.SVD++是当前最为常用的协同过滤模型之一,差分隐私模型则是近十年来隐私保护理论最重要的研究进展之一,将两者相结合提出3种基于差分隐私和SVD++的协同过滤模型:基于梯度扰动的SVD++隐私保护模型、基于目标函数扰动的SVD++隐私保护模型和基于输出结果扰动的隐私保护模型.理论分析和实验结果显示,所提出的算法不仅能为用户的隐私安全提供可靠的保障,而且还可保持较高的预测准确度.     

基于评分矩阵局部低秩假设的成列协同排名算法

协同过滤方法是当今大多数推荐系统的核心.传统的协同过滤方法专注于评分预测的准确性,然而实际推荐系统的推荐结果往往是项目的排序.针对这一问题,将排名学习领域的知识引入推荐算法,设计了一种基于评分矩阵局部低秩假设的成列协同排名算法.选择直接使用计算复杂度较低的成列损失函数来优化矩阵分解模型,并通过实验验证了其在运算速度上的显著提升.在3个实际推荐系统数据集上,与当下主流推荐算法的比较实验结果表明,该算法具有良好的性能.     

面向用户偏好建模的个性化对话推荐算法

推荐系统的目标是从物品数据库中，选择出与用户兴趣偏好相匹配的子集，缓解用户面临的“信息过载”问题。因而近年来推荐系统越来越多地应用到电商、社交等领域，展现出巨大的商业潜力。传统推荐系统中，系统对用户的认知往往来源于历史交互记录，例如点击率或者购买记录，这是一种隐式用户反馈。对话推荐系统能够通过自然语言与用户进行多轮对话，逐步深入挖掘其兴趣偏好，从而向对方提供高质量的推荐结果。相比于传统推荐系统，对话推荐系统主要有两方面的不同。其一，对话推荐系统能够利用自然语言与用户进行语义上连贯的多轮对话，提升了人机交互中的用户体验；其二，系统能够询问特定的问题直接获取用户的显式反馈，从而更深入地理解用户兴趣偏好，提供更可靠的推荐结果。目前已经有不少工作在不同的问题设定下对该领域进行了探索，然而尽管如此，这些工作仍仅局限于关注当前正在进行的对话，忽视了过去交互记录中蕴涵的丰富信息，导致对用户偏好建模的不充分。为了解决这个问题，本文提出了一个面向用户偏好建模的个性化对话推荐算法框架，通过双线性模型注意力机制与自注意力层次化编码结构进行用户偏好建模，从而完成对候选物品的排序与推荐。本文设计的模型结构能够在充分利用用户历史对话信息的同时，权衡历史对话与当前对话两类数据的重要性。丰富的用户相关信息来源使得推荐结果在契合用户个性化偏好的同时，更具备多样性，从而缓解“信息茧房”等现象带来的不良影响。基于公开数据集的实验表明了本文方法在个性化对话推荐任务上的有效性。     

A Multi-dimensional Trust-aware Cloud Service Selection Mechanism Based on Evidential Reasoning Approach

In the last few years, cloud computing as a new computing paradigm has gone through significant development, but it is also facing many problems. One of them is the cloud service selection problem. As increasingly boosting cloud services are offered through the internet and some of them may be not reliable or even malicious, how to select trustworthy cloud services for cloud users is a big challenge. In this paper, we propose a multi-dimensional trust-aware cloud service selection mechanism based on evidential reasoning(ER) approach that integrates both perception-based trust value and reputation based trust value, which are derived from direct and indirect trust evidence respectively, to identify trustworthy services. Here, multi-dimensional trust evidence, which reflects the trustworthiness of cloud services from different aspects, is elicited in the form of historical users feedback ratings. Then, the ER approach is applied to aggregate the multi-dimensional trust ratings to obtain the real-time trust value and select the most trustworthy cloud service of certain type for the active users. Finally, the fresh feedback from the active users will update the trust evidence for other service users in the future.     

A Multicriteria Collaborative Filtering Approach Using the Indifference Relation and its Application to Initiator Recommendation for Group-Buying

For recommender systems, the main aim of the popular collaborative filtering approaches is to recommend items that users with similar preferences have liked in the past. Single-criterion recommender systems have been successfully used in several applications. Because leveraging multicriteria information can potentially improve recommendation accuracy, multicriteria rating systems that allow users to assign ratings to various content attributes of items they have consumed have become the focus in recommendation systems. By treating the recommendation of items as a multicriteria decision problem, it is interesting to incorporate the preference relation of users of multicriteria decision making (MCDM) into the similarity measure for a collaborative filtering approach. For this, the well-known indifference relation can justify a discrimination or similarity between any two users, if outranking relation theory is incorporated. The applicability of the proposed single-criterion and multicriteria recommendation approaches to the recommendation of initiators on a group-buying website was examined. Experimental results have demonstrated that the generalization ability of the proposed multicriteria recommendation approach performs well in comparison to other single-criterion and multicriteria collaborative filtering approaches.     

Trustworthy assurance of service interoperation in cloud environment

Cloud computing can be realized by service interoperation and its essence is to provide cloud services through network. The development of effective methods to assure the trustworthiness of service interoperation in cloud environment is a very important problem. The essence of cloud security is trust and trust management. Combining quality of service (QoS) with trust model, this paper constructs a QoS-aware and quantitative trust-model that consists of initial trust value, direct trust value, and recommendatory trust value of service, making the provision, discovery, and aggregation of cloud services trustworthy. Hence, it can assure trustworthiness of service interoperation between users and services or among services in cloud environment. At the same time, based on this model, service discovery method based on QoS-aware and quantitative trust-model (TQoS-WSD) is proposed, which makes a solid trust relationship among service requestor, service provider and service recommender, and users can find trustworthy service whose total evaluation value is higher. Compared to QoS-based service discovery (QoS-WSD) method, it is proved by the experiment for TQoS-WSD method that more accurate result of service discovery will be achieved by service requestor, while reasonable time cost is increased. Meanwhile, TQoS-WSD method strongly resists the effect of service discovery by untrustworthy QoS values and improves service invocation success-rate and thus assures trustworthiness of services interoperation.     

Semantic inference of user’s reputation and expertise to improve collaborative recommendations

Collaborative recommender systems select potentially interesting items for each user based on the preferences of like-minded individuals. Particularly, e-commerce has become a major domain in these research field due to its business interest, since identifying the products the users may like or find useful can boost consumption. During the last years, a great number of works in the literature have focused in the improvement of these tools. Expertise, trust and reputation models are incorporated in collaborative recommender systems to increase their accuracy and reliability. However, current approaches require extra data from the users that is not often available. In this paper, we present two contributions that apply a semantic approach to improve recommendation results transparently to the users. On the one hand, we automatically build implicit trust networks in order to incorporate trust and reputation in the selection of the set of like-minded users that will drive the recommendation. On the other hand, we propose a measure of practical expertise by exploiting the data available in any e-commerce recommender system – the consumption histories of the users.     

融合信任关系和有用性评价的矩阵分解推荐方法

冷启动和数据稀疏性问题是推荐系统面临的两大难题。现有的大多数基于矩阵分解的推荐方法将用户孤立对待,忽略了用户之间的信任关系,导致推荐性能较低。提出一种融合信任关系和有用性评价的矩阵分解推荐方法。该方法在对评分矩阵进行概率分解的基础上,加入有用性评价和用户信任关系,采用交替最小二乘法训练模型参数。Epinions和Ciao数据集上的对比实验表明,所提方法有效提高了推荐系统的准确性和可靠性,尤其存在冷启动用户时,该方法的推荐精度明显优于传统的推荐方法。     

基于信任网络随机游走模型的协同过滤推荐

协同过滤是目前应用最广泛和最成功的推荐技术之一。然而,目前该技术的发展面临着严重的冷启动和稀疏性问题,降低了其推荐质量,因此提出了一种基于信任网络随机游走模型的协同过滤推荐方法。该方法融合了基于信任和项目的协同过滤推荐方法,并引入了信任因子作为引导推荐的重要因素。随机游走模型不仅考虑了信任用户对目标项目的评分,也考虑了他们对与目标项目相似的项目的评分。随着随机游走深度的增加,以相似项目的评分信息来代替目标项目的评分信息的概率也逐渐增大。在Epinions真实数据集上的验证结果表明,该方法在推荐评价指标上比其他算法具有更好的推荐结果。     

耦合辅助信息的矩阵分解推荐模型

近十年来，协同过滤（CF）推荐系统成功地为用户提供了个性化的产品和服务。然而，用户—物品矩阵的稀疏性、推荐精度不高等问题仍然是一个挑战。针对这些问题，在矩阵分解模型基础上，提出了耦合用户和物品辅助信息的矩阵分解混合协同过滤框架；然后，基于此框架又提出了耦合物品属性信息相似度（COS）的过滤模型。大规模真实数据集上的实验表明，该模型不但可以有效解决物品相似度度量问题，而且相比传统方法，尤其是在物品特征非常稀疏的情况下，推荐准确性得到有效改进。     

结合信任的推荐系统的性质

结合信任的推荐系统可以有效地缓解传统协同过滤算法中存在的数据稀疏问题,并能给每个用户提供可信且准确的推荐。然而系统中的每个用户都是不同的,因此考虑针对不同用户应采用不同推荐模式来查找推荐群体,以做出更具个性化的推荐。研究了微观层次上的节点特性,引入了兴趣的概念,证明了被推荐者的多种节点特性对于推荐结果的影响效果。最后通过多组实验验证了推荐系统在具有不同特性的节点上的推荐效果差异。