Distributed collaborative filtering with singular ratings for large scale recommendation

Collaborative filtering (CF) is an effective technique addressing the information overloading problem, where each user is associated with a set of rating scores on a set of items. For a chosen target user, conventional CF algorithms measure similarity between this user and other users by utilizing pairs of rating scores on common rated items, but discarding scores rated by one of them only. We call these comparative scores as dual ratings, while the non-comparative scores as singular ratings. Our experiments show that only about 10% ratings are dual ones that can be used for similarity evaluation, while the other 90% are singular ones. In this paper, we propose SingCF approach, which attempts to incorporate multiple singular ratings, in addition to dual ratings, to implement collaborative filtering, aiming at improving the recommendation accuracy. We first estimate the unrated scores for singular ratings and transform them into dual ones. Then we perform a CF process to discover neighborhood users and make predictions for each target user. Furthermore, we provide a MapReduce-based distributed framework on Hadoop for significant improvement in efficiency. Experiments in comparison with the state-of-the-art methods demonstrate the performance gains of our approaches.     

An algorithm for efficient privacy-preserving item-based collaborative filtering

Collaborative filtering (CF) methods are widely adopted by existing recommender systems, which can analyze and predict user “ratings” or “preferences” of newly generated items based on user historical behaviors. However, privacy issue arises in this process as sensitive user private data are collected by the recommender server. Recently proposed privacy-preserving collaborative filtering (PPCF) methods, using computation-intensive cryptography techniques or data perturbation techniques are not appropriate in real online services. In this paper, an efficient privacy-preserving item-based collaborative filtering algorithm is proposed, which can protect user privacy during online recommendation process without compromising recommendation accuracy and efficiency. The proposed method is evaluated using the Netflix Prize dataset. Experimental results demonstrate that the proposed method outperforms a randomized perturbation based PPCF solution and a homomorphic encryption based PPCF solution by over 14X and 386X, respectively, in recommendation efficiency while achieving similar or even better recommendation accuracy.     

On the combination of user-based and item-based collaborative filtering

In this paper, we propose two new filtering algorithms which are a combination of user-based and item-based collaborative filtering schemes. The first one, Hybrid-Ib, identifies a reasonably large neighbourhood of similar users and then uses this subset to derive the item-based recommendation model. The second algorithm, Hybrid-CF, starts by locating items similar to the one for which we want a prediction, and then, based on that neighbourhood, it generates its user-based predictions. We start by describing the execution steps of the algorithms and proceed with extended experiments. We conclude that our algorithms are directly comparable to existing filtering approaches, with Hybrid-CF producing favorable or, in the worst case, similar results in all selected evaluation metrics.     

Property-based collaborative filtering for health-aware recommender systems

Recommender systems aim at solving the problem of information overload by selecting items (commercial products, educational assets, TV programs, etc.) that match the consumers’ interests and preferences. Recently, there have been approaches to drive the recommendations by the information stored in electronic health records, for which the traditional strategies applied in online shopping, e-learning, entertainment and other areas have several pitfalls. This paper addresses those problems by introducing a new filtering strategy, centered on the properties that characterize the items and the users. Preliminary experiments with real users have proved that this approach outperforms previous ones in terms of consumers’ satisfaction with the recommended items. The benefits are especially apparent among people with specific health concerns.     

The impact of data obfuscation on the accuracy of collaborative filtering

Collaborative filtering (CF) is a widely-used technique for generating personalized recommendations. CF systems are typically based on a central storage of user profiles, i.e., the ratings given by users to items. Such centralized storage introduces potential privacy breach, since all the user profiles may be accessible by untrusted parties when breaking the access control of the centralized system. Hence, recent studies have focused on enhancing the privacy of CF users by distributing their user profiles across multiple repositories and obfuscating the user profiles to partially hide the actual user ratings. This work combines these two techniques and investigates the unavoidable side effect of data obfuscation: the reduction of the accuracy of the generated CF predictions. The evaluation, which was conducted using three different datasets, shows that considerable parts of the user profiles can be modified without observing a substantial decrease of the CF prediction accuracy. The evaluation also indicates what parts of the user profiles are required for generating accurate CF predictions. In addition, we conducted an exploratory user study that reveals positive attitude of users towards the data obfuscation.     

A novel collaborative filtering approach for recommending ranked items

In recent years, Collaborative Filtering (CF) has proven to be one of the most successful techniques used in recommendation systems. Since current CF systems estimate the ratings of not-yet-rated items based on other items’ ratings, these CF systems fail to recommend products when users’ preferences are not expressed in numbers. In many practical situations, however, users’ preferences are represented by ranked lists rather than numbers, such as lists of movies ranked according to users’ preferences. Therefore, this study proposes a novel collaborative filtering methodology for product recommendation when the preference of each user is expressed by multiple ranked lists of items. Accordingly, a four-staged methodology is developed to predict the rankings of not-yet-ranked items for the active user. Finally, a series of experiments is performed, and the results indicate that the proposed methodology produces high-quality recommendations.     

A collaborative filtering framework based on fuzzy association rules and multiple-level similarity

The rapid development of Internet technologies in recent decades has imposed a heavy information burden on users. This has led to the popularity of recommender systems, which provide advice to users about items they may like to examine. Collaborative Filtering (CF) is the most promising technique in recommender systems, providing personalized recommendations to users based on their previously expressed preferences and those of other similar users. This paper introduces a CF framework based on Fuzzy Association Rules and Multiple-level Similarity (FARAMS). FARAMS extended existing techniques by using fuzzy association rule mining, and takes advantage of product similarities in taxonomies to address data sparseness and nontransitive associations. Experimental results show that FARAMS improves prediction quality, as compared to similar approaches. Cane Wing-ki Leung is a PhD student in the Department of Computing, The Hong Kong Polytechnic University, where she received her BA degree in Computing in 2003. Her research interests include collaborative filtering, data mining and computer-supported collaborative work. Stephen Chi-fai Chan is an Associate Professor and Associate Head of the Department of Computing, The Hong Kong Polytechnic University. Dr. Chan received his PhD from the University of Rochester, USA, worked on computer-aided design at Neo-Visuals, Inc. in Toronto, Canada, and researched in computer-integrated manufacturing at the National Research Council of Canada before joining the Hong Kong Polytechnic University in 1993. He is currently working on the development of collaborative Web-based information systems, with applications in education, electronic commerce, and manufacturing. Fu-lai Chung received his BSc degree from the University of Manitoba, Canada, in 1987, and his MPhil and PhD degrees from the Chinese University of Hong Kong in 1991 and 1995, respectively. He joined the Department of Computing, Hong Kong Polytechnic University in 1994, where he is currently an Associate Professor. He has published widely in the areas of computational intelligence, pattern recognition and recently data mining and multimedia in international journals and conferences and his current research interests include time series data mining, Web data mining, bioinformatics data mining, multimedia content analysis,and new computational intelligence techniques.     

Cluster ensembles in collaborative filtering recommendation

Recommender systems, which recommend items of information that are likely to be of interest to the users, and filter out less favored data items, have been developed. Collaborative filtering is a widely used recommendation technique. It is based on the assumption that people who share the same preferences on some items tend to share the same preferences on other items. Clustering techniques are commonly used for collaborative filtering recommendation. While cluster ensembles have been shown to outperform many single clustering techniques in the literature, the performance of cluster ensembles for recommendation has not been fully examined. Thus, the aim of this paper is to assess the applicability of cluster ensembles to collaborative filtering recommendation. In particular, two well-known clustering techniques (self-organizing maps (SOM) and k-means), and three ensemble methods (the cluster-based similarity partitioning algorithm (CSPA), hypergraph partitioning algorithm (HGPA), and majority voting) are used. The experimental results based on the Movielens dataset show that cluster ensembles can provide better recommendation performance than single clustering techniques in terms of recommendation accuracy and precision. In addition, there are no statistically significant differences between either the three SOM ensembles or the three k-means ensembles. Either the SOM or k-means ensembles could be considered in the future as the baseline collaborative filtering technique.     

A hybrid content-based and item-based collaborative filtering approach to recommend TV programs enhanced with singular value decomposition

With the advent of new cable and satellite services, and the next generation of digital TV systems, people are faced with an unprecedented level of program choice. This often means that viewers receive much more information than they can actually manage, which may lead them to believe that they are missing programs that could likely interest them. In this context, TV program recommendation systems allow us to cope with this problem by automatically matching user’s likes to TV programs and recommending the ones with higher user preference.This paper describes the design, development, and startup of queveo.tv: a Web 2.0 TV program recommendation system. The proposed hybrid approach (which combines content-filtering techniques with those based on collaborative filtering) also provides all typical advantages of any social network, such as supporting communication among users as well as allowing users to add and tag contents, rate and comment the items, etc. To eliminate the most serious limitations of collaborative filtering, we have resorted to a well-known matrix factorization technique in the implementation of the item-based collaborative filtering algorithm, which has shown a good behavior in the TV domain. Every step in the development of this application was taken keeping always in mind the main goal: to simplify as much as possible the user task of selecting what program to watch on TV.     

Pessimists and optimists: Improving collaborative filtering through sentiment analysis

This work presents a novel application of Sentiment Analysis in Recommender Systems by categorizing users according to the average polarity of their comments. These categories are used as attributes in Collaborative Filtering algorithms. To test this solution a new corpus of opinions on movies obtained from the Internet Movie Database (IMDb) has been generated, so both ratings and comments are available. The experiments stress the informative value of comments. By applying Sentiment Analysis approaches some Collaborative Filtering algorithms can be improved in rating prediction tasks. The results indicate that we obtain a more reliable prediction considering only the opinion text (RMSE of 1.868), than when apply similarities over the entire user community (RMSE of 2.134) and sentiment analysis can be advantageous to recommender systems.     

A domain model of Web recommender systems based on usage mining and collaborative filtering

Considering the increasing demand of multi-agent systems, the practice of software reuse is essential to the development of such systems. Multi-agent domain engineering is a process for the construction of domain-specific agent-based reusable software artifacts, like domain models, representing the requirements of a family of multi-agent systems in a domain, and frameworks, implementing reusable agent-based design solutions to those requirements. This article describes the domain modeling tasks of the MADEM methodology and a case study on the application of GRAMO, a MADEM technique, for the construction of the domain model of ONTOWUM, specifying the common and variable requirements of a family of Web recommender systems based on usage mining and collaborative filtering.     

Userrank for item-based collaborative filtering recommendation

With the recent explosive growth of the Web, recommendation systems have been widely accepted by users. Item-based Collaborative Filtering (CF) is one of the most popular approaches for determining recommendations. A common problem of current item-based CF approaches is that all users have the same weight when computing the item relationships. To improve the quality of recommendations, we incorporate the weight of a user, userrank, into the computation of item similarities and differentials. In this paper, a data model for userrank calculations, a PageRank-based user ranking approach, and a userrank-based item similarities/differentials computing approach are proposed. Finally, the userrank-based approaches improve the recommendation results of the typical Adjusted Cosine and Slope One item-based CF approaches.     

Collaborative filtering based on workflow space

The traditional recommender systems are usually oriented to general situations in daily lives (e.g. recommend movies, books, music, news and etc.), but seldom cover the recommendation scenarios for the collaborative team environments. We have done an explorative study on collaborative filtering mechanism for collaborative team environments, which is some kind of multi-dimensional recommender systems problem with consideration of workflow context. This paper proposed 3-dimensional workflow space model, and investigated the new similarities measure between members in workflow space. Then, the new similarities measure is utilized into collaborative filtering for recommender systems in collaborative team environments. At last, the efficiency and usability of the proposed method are validated by experiments referring to a real-world collaborative team of a manufacturing enterprise.     

Enhancing the diversity of conversational collaborative recommendations: a comparison

In conversational collaborative recommender systems, user feedback influences the recommendations. We report mechanisms for enhancing the diversity of the recommendations made by collaborative recommenders. We focus on techniques for increasing diversity that rely on collaborative data only. In our experiments, we compare different mechanisms and show that, if recommendations are diverse, users find target items in many fewer recommendation cycles.     

基于用户评论的动态方面注意力电商推荐深度学习模型

为了提供个性化推荐,推荐系统会将用户和物品分别表达为用户偏好向量和物品特征向量。物品特征向量中不同维度分别对应物品不同的特征。用户偏好向量中各维度表示用户对物品对应维度(特征)的喜好程度。目前大部分的推荐算法都假设为对于不同物品、同一用户的偏好向量是相同的。然而在现实生活中,该假设是不成立的。为此,提出一种结合注意力机制的深度学习模型,其能根据不同的用户-物品对,相应地学习到一个注意力权重向量,最终达到动态调整用户偏好向量的目的。在3组公开数据集上进行对比实验,以预测评分的均方误差(MSE)作为评估指标,实验结果表明该方法比已有的相关算法的效果更好。     

基于项目分类的协同过滤改进算法*

为了解决用户评分数据稀疏性和用户最近邻寻找的准确性问题,提出了一种基于项目分类的协同过滤推荐改进算法。该算法首先利用项目分类信息为类内未评分项目预测评分值;然后通过计算类内用户间的相似度得到目标用户的最近邻居;最后进行推荐。实验结果表明,该算法可以准确地获取用户兴趣最近邻,有效地解决数据稀疏性问题;同时,该算法还极大地提高了系统的工作效率及可扩展性。     

一种基于Sigmoid函数的改进协同过滤推荐算法

随着电子商务和社交网络的蓬勃发展, 推荐系统逐渐成为数据挖掘领域的重要研究方向。推荐系统能够从海量信息中定位用户兴趣点, 提供个性化服务。协同过滤算法能够有效分析用户偏好, 提供合适的推荐服务。针对评分矩阵稀疏时传统协同过滤算法性能很差的问题, 提出一种基于Sigmoid函数的改进推荐系统算法。利用Sigmoid函数对不同项目进行建模, 得到项目的平均受欢迎程度; 利用Sigmoid函数对不同用户进行建模, 将评分映射为用户对项目的喜好程度; 根据用户对项目喜好程度应该与项目平均受欢迎程度贴近的原则进行评分预测。在两组真实数据集合上的实验结果表明, 该算法较好地解决了数据稀疏性问题, 能够有效提高传统算法的预测准确性。     

自适应用户的Item-based协同过滤推荐算法

传统Item-based协同过滤算法计算两个条目间相似性时, 将每个评分视为同等重要, 忽略了共评用户（对两个条目共同评分的用户）与目标用户间的相似性对条目间相似性的影响。针对此问题, 提出了一种自适应用户的Item-based协同过滤算法。该算法将共评用户与目标用户的相似性作为共评用户评分重要性的权重, 以实现针对不同的目标用户, 为目标条目选择不同的、适合目标用户的最近邻居集, 从而提高推荐准确性。实验结果表明, 提出的算法可以显著提高推荐系统的推荐质量。     

Improving memory-based collaborative filtering via similarity updating and prediction modulation

Memory-based collaborative filtering (CF) makes recommendations based on a collection of user preferences for items. The idea underlying this approach is that the interests of an active user will more likely coincide with those of users who share similar preferences to the active user. Hence, the choice and computation of a similarity measure between users is critical to rating items. This work proposes a similarity update method that uses an iterative message passing procedure. Additionally, this work deals with a drawback of using the popular mean absolute error (MAE) for performance evaluation, namely that ignores ratings distribution. A novel modulation method and an accuracy metric are presented in order to minimize the predictive accuracy error and to evenly distribute predicted ratings over true rating scales. Preliminary results show that the proposed similarity update and prediction modulation techniques significantly improve the predicted rankings.