This paper presents an overview of the field of recommender systems and describes the current generation of recommendation methods that are usually classified into the following three main categories: content-based, collaborative, and hybrid recommendation approaches. This paper also describes various limitations of current recommendation methods and discusses possible extensions that can improve recommendation capabilities and make recommender systems applicable to an even broader range of applications. These extensions include, among others, an improvement of understanding of users and items, incorporation of the contextual information into the recommendation process, support for multcriteria ratings, and a provision of more flexible and less intrusive types of recommendations.

Recommender Systems (RS) aim at predicting items or ratings of items that the user are interested in. Collaborative Filtering (CF) algorithms such as user- and item-based methods are the dominant techniques applied in RS algorithms. To improve recommendation quality, metadata such as content information of items has typically been used as additional knowledge. With the increasing popularity of the collaborative tagging systems, tags could be interesting and useful information to enhance RS algorithms. Unlike attributes which are “global ” descriptions of items, tags are “local ” descriptions of items given by the users. To the best of our knowledge, there hasn’t been any prior study on tagaware RS. In this paper, we propose a generic method that allows tags to be incorporated to standard CF algorithms, by reducing the three-dimensional correlations to three twodimensional correlations and then applying a fusion method to re-associate these correlations. Additionally, we investigate the effect of incorporating tags information to different CF algorithms. Empirical evaluations on three CF algorithms with real-life data set demonstrate that incorporating tags to our proposed approach provides promising and significant results.

Abstract. Tagging systems (or folksonomies) like Flickr or Delicious are expanding tremendously. More and more resources are being added to them. As the resources present on these system increase in amount, it becomes difficult to explore these resources. For this purpose, we present a system T-ORG, which provides a mechanism to organize these resources by classifying the tags (or keywords) attached to them into predefined categories. Supervised classification in this case seems infeasible; therefore we also propose a new classification algorithm T-KNOW that does not require training data. For our experiments, we have downloaded images and their tags from groups present on Flickr website and then classified these tags into different categories. We have used Cohen’s Kappa and F-measure to evaluate the classification results of T-KNOW. Results are encouraging and show that T-ORG can be used to explore resources in an effective manner.

Recent studies have indicated that the application of Multi-Criteria Decision Making (MCDM) methods in recommender systems has yet to be systematically explored. This observation partially contradicts with the fact that in related literature, there exist several contributions describing recommender systems that engage some MCDM method. Such systems, which we refer to as multi-criteria recommender systems, have early demonstrated the potential of applying MCDM methods to facilitate recommendation, in numerous application domains. On the other hand, a comprehensive analysis of existing systems would facilitate their understanding and development. Towards this direction, this paper identifies a set of dimensions that distinguish, describe and categorize multi-criteria recommender systems, based on existing taxonomies and categorizations. These dimensions are integrated into an overall framework that is used for the analysis and classification of a sample of existing multi-criteria recommender systems. The results provide a comprehensive overview of the ways current multi-criteria recommender systems support the decision of online users.

Abstract: The need to support users of the Internet with the selection of information is becoming more important. Learners in complex, self-organising Learning Networks have similar problems and need guidance to find and select most suitable learning activities, in order to attain their lifelong learning goals in the most efficient way. Several research questions regarding efficiency and effectiveness deal with adequate navigation support through recommender systems. To answer some of these questions an experiment was set up within an Introduction Psychology course of the Open University of the Netherlands. Around 250 students participated in this study and were monitored over an experimental period of four months. All were provided the same course materials, but only half of them were supported with a personalised recommender system. This study examined the effects of the navigation

The paper presents a method that uses aggregate ratings provided by various segments of users for various categories of items to derive better estimations of unknown individual ratings. This is achieved by converting the aggregate ratings into constraints on the parameters of a rating estimation model presented in the paper. The paper also demonstrates theoretically that these additional constraints reduce rating estimation errors resulting in better rating predictions.

Abstract not found

The situation in which a choice is made is an important information for recommender systems. Context-aware recommenders take this information into account to make predictions. So far, the best performing method for contextaware rating prediction in terms of predictive accuracy is Multiverse Recommendation based on the Tucker tensor factorization model. However this method has two drawbacks: (1) itsmodel complexityisexponentialinthenumberofcontext variables and polynomial in the size of the factorization and (2) it only works for categorical context variables. On the other hand there is a large variety of fast but specialized recommender methods which lack the generality of contextaware methods. We propose to apply Factorization Machines (FMs) to model contextual information and to provide context-aware rating predictions. This approach results in fast contextaware recommendations because the model equation of FMs can be computed in linear time both in the number of context variables and the factorization size. For learning FMs, we develop an iterative optimization method that analytically finds the least-square solution for one parameter given the other ones. Finally, we show empirically that our approach outperforms Multiverse Recommendation in prediction quality and runtime. Categories andSubject Descriptors

The Semantic Logger 1 (SL) is presented as a system for the importing, housing, and exploiting of personal information. The system has been implemented using a number of Semantic Web enabling technologies, and attempts to store the information in a manner adhering to as many W3C recommendations as possible. The Semantic Logger’s utility is grounded in two context-based applications, namely a recommender system, and a photo-annotation tool. Categories and Subject Descriptors

The current Web service architecture addressed service discovery problem, but not service selection. If we treat services as special type of products from the service providers, existing techniques used for product selection can be applied to service selection. Among many existing techniques for product selection, one dominant approach is to use recommender systems. For example many ecommerce Web site, such as Ebay, Amazon and Epinions all have recommender system support to ease the burden of product selection. Recommender systems are classified into different types, which include content-based, collaborative and hybrid based systems which perform recommendation based on user/item data set. There is also multidimensional recommender system where the system supports multiple dimensions, such as user,item, and its quality attributes. This research investigates the fundamentals of different types recommender systems. In particular, the collaborative filtering based approach using both two dimensional and multidimensional data. This study is to facilitate the development