HCI Bibliography Home | HCI Conferences | RecSys Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
RecSys Tables of Contents: 070809101112131415

Proceedings of the 2009 ACM Conference on Recommender Systems

Fullname:Proceedings of the Third ACM Conference on Recommender Systems
Editors:Lawrence Bergman; Alex Tuzhilin; Robin Burke; Alexander Felfernig; Lars Schmidt-Thieme
Location:New York, New York
Dates:2009-Oct-23 to 2009-Oct-25
Standard No:ISBN: 1-60558-435-5, 978-1-60558-435-5; ACM DL: Table of Contents hcibib: RecSys09
Links:Conference Home Page
  1. Algorithms I
  2. Tags and social networks
  3. Applications
  4. Algorithms II
  5. Privacy and security
  6. Trust and evaluation
  7. Short papers
  8. Doctoral symposium
  9. Workshops
  10. Tutorials
Recsys'09 industrial keynote: top 10 lessons learned developing deploying and operating real-world recommender systems BIBAFull-Text 1-2
  Francisco J. Martin
The number of online services providing users with real-time recommendations has increased exponentially over the last few years. Recommender Systems that were originally only accessible to a limited number of high-tech companies are now widely available through a growing number of both technical choices and vendors. The acceptance however, of automatically delivered recommendations by users depends on numerous factors that go far beyond the algorithms that constitute the major focus of researchers. This talk will share a number of lessons learned over the last ten years creating and operating recommender systems in a multitude of domains, from digital music to fitness plans through personal finance management, and in a multitude of business settings, from lightweight integrations to highly-coupled integrations within secure bank environments.
Up close and personalized: a marketing view of recommendation systems BIBAFull-Text 3-4
  Michel Wedel; Roland T. Rust; Tuck Siong Chung
Developments in the marketing literature on recommendation systems are reviewed and an illustration of an Adaptive Personalization System is provided in the context of music. This illustration reveals that Adaptive Personalization Systems have the potential to significantly increase the effectiveness of personal recommendations, and perform better than extant methods.
Collaborative prediction and ranking with non-random missing data BIBAFull-Text 5-12
  Benjamin M. Marlin; Richard S. Zemel
A fundamental aspect of rating-based recommender systems is the observation process, the process by which users choose the items they rate. Nearly all research on collaborative filtering and recommender systems is founded on the assumption that missing ratings are missing at random. The statistical theory of missing data shows that incorrect assumptions about missing data can lead to biased parameter estimation and prediction. In a recent study, we demonstrated strong evidence for violations of the missing at random condition in a real recommender system. In this paper we present the first study of the effect of non-random missing data on collaborative ranking, and extend our previous results regarding the impact of non-random missing data on collaborative prediction.

Algorithms I

A spatio-temporal approach to collaborative filtering BIBAFull-Text 13-20
  Zhengdong Lu; Deepak Agarwal; Inderjit S. Dhillon
In this paper, we propose a novel spatio-temporal model for collaborative filtering applications. Our model is based on low-rank matrix factorization that uses a spatio-temporal filtering approach to estimate user and item factors. The spatial component regularizes the factors by exploiting correlation across users and/or items, modeled as a function of some implicit feedback (e.g., who rated what) and/or some side information (e.g., user demographics, browsing history). In particular, we incorporate correlation in factors through a Markov random field prior in a probabilistic framework, whereby the neighborhood weights are functions of user and item covariates. The temporal component ensures that the user/item factors adapt to process changes that occur through time and is implemented in a state space framework with fast estimation through Kalman filtering. Our spatio-temporal filtering (ST-KF hereafter) approach provides a single joint model to simultaneously incorporate both spatial and temporal structure in ratings and therefore provides an accurate method to predict future ratings. To ensure scalability of ST-KF, we employ a mean-field approximation for inference. Incorporating user/item covariates in estimating neighborhood weights also helps in dealing with both cold-start and warm-start problems seamlessly in a single unified modeling framework; covariates predict factors for new users and items through the neighborhood. We illustrate our method on simulated data, benchmark data and data obtained from a relatively new recommender system application arising in the context of Yahoo! Front Page.
Pairwise preference regression for cold-start recommendation BIBAFull-Text 21-28
  Seung-Taek Park; Wei Chu
Recommender systems are widely used in online e-commerce applications to improve user engagement and then to increase revenue. A key challenge for recommender systems is providing high quality recommendation to users in ``cold-start" situations. We consider three types of cold-start problems: 1) recommendation on existing items for new users; 2) recommendation on new items for existing users; 3) recommendation on new items for new users. We propose predictive feature-based regression models that leverage all available information of users and items, such as user demographic information and item content features, to tackle cold-start problems. The resulting algorithms scale efficiently as a linear function of the number of observations. We verify the usefulness of our approach in three cold-start settings on the MovieLens and EachMovie datasets, by comparing with five alternatives including random, most popular, segmented most popular, and two variations of Vibes affinity algorithm widely used at Yahoo! for recommendation.
Ordering innovators and laggards for product categorization and recommendation BIBAFull-Text 29-36
  Sarah K. Tyler; Shenghuo Zhu; Yun Chi; Yi Zhang
Different buyers exhibit different purchasing behaviors. Some rush to purchase new products while others tend to be more cautious, waiting for reviews from people they trust. In market analysis, the former group of buyers is often referred to as innovators and early adopters while the latter group is referred to as laggards. The adoption behavior is a dynamic feature of the user and varies over groups of products, e.g., innovators of literature may not be the innovators of electronics. The adoption order of users is a dynamic feature of the product, which can help to predict the future potential buyers. However, such dynamic features are usually unavailable in the description of products. In this paper, we study the user behavior of an online review website -- Epinions.com. We first propose to model user adoption behaviors by creating a total ordering among users who rate the products in a given category. We develop a greedy algorithm and a Markov-chain based algorithm for computing the category total ordering. Next, we show that by using user behavior information, we can more accurately predict the category of a new product as well as predict which users will follow. Furthermore, by using the Epinion.com trust network as evidence, we demonstrate that our total ordering can group users into communities that closely resemble the trust network. Thus the adoption order can be a useful feature in recommendation systems.
Improving rating estimation in recommender systems using aggregation- and variance-based hierarchical models BIBAFull-Text 37-44
  Akhmed Umyarov; Alexander Tuzhilin
Previous work on using external aggregate rating information showed that this information can be incorporated in several different types of recommender systems and improves their performance. In this paper, we propose a more general class of methods that combine external aggregate information with individual ratings in a novel way. Unlike the previously proposed methods, one of the defining features of this approach is that it takes into the consideration not only the aggregate average ratings but also the variance of the aggregate distribution of ratings. The methods proposed in this paper estimate unknown ratings by finding an optimal linear combination of individual-level and aggregate-level rating estimators in a form of a hierarchical regression (HR) model that is grounded in the theory of statistics and machine learning.
   The proposed HR model is general enough so that the standard individual-level recommender systems and naive aggregate methods constitute special cases of this model.
   We show that for the general HR model, the presence of the aggregate variance, surprisingly, does not significantly improve estimation of unknown ratings vis-a-vis the case when only aggregate average ratings are considered.
   In the paper, we experimentally show that the optimal linear combination approach significantly dominates all other special cases, including the classical non-aggregated case and our previously studied aggregate methods, and therefore is the method of choice.

Tags and social networks

The impact of ambiguity and redundancy on tag recommendation in folksonomies BIBAFull-Text 45-52
  Jonathan Gemmell; Maryam Ramezani; Thomas Schimoler; Laura Christiansen; Bamshad Mobasher
Collaborative tagging applications have become a popular tool allowing Internet users to manage online resources with tags. Most collaborative tagging applications permit unsupervised tagging resulting in tag ambiguity in which a single tag has many different meanings and tag redundancy in which several tags have the same meaning. Common metrics for evaluating tag recommenders may overestimate the utility of ambiguous tags or ignore the appropriateness of redundant tags. Ambiguity and redundancy may even burden the user with additional effort by requiring them to clarify an annotation or forcing them to distinguish between highly related items. In this paper we demonstrate that ambiguity and redundancy impede the evaluation and performance of tag recommenders. Five tag recommendation strategies based on popularity, collaborative filtering and link analysis are explored. We use a cluster-based approach to define ambiguity and redundancy and provide extensive evaluation on three real world datasets.
Personalized recommendation of social software items based on social relations BIBAFull-Text 53-60
  Ido Guy; Naama Zwerdling; David Carmel; Inbal Ronen; Erel Uziel; Sivan Yogev; Shila Ofek-Koifman
We study personalized recommendation of social software items, including bookmarked web-pages, blog entries, and communities. We focus on recommendations that are derived from the user's social network. Social network information is collected and aggregated across different data sources within our organization. At the core of our research is a comparison between recommendations that are based on the user's familiarity network and his/her similarity network. We also examine the effect of adding explanations to each recommended item that show related people and their relationship to the user and to the item. Evaluation, based on an extensive user survey with 290 participants and a field study including 90 users, indicates superiority of the familiarity network as a basis for recommendations. In addition, an important instant effect of explanations is found -- interest rate in recommended items increases when explanations are provided.
Latent dirichlet allocation for tag recommendation BIBAFull-Text 61-68
  Ralf Krestel; Peter Fankhauser; Wolfgang Nejdl
Tagging systems have become major infrastructures on the Web. They allow users to create tags that annotate and categorize content and share them with other users, very helpful in particular for searching multimedia content. However, as tagging is not constrained by a controlled vocabulary and annotation guidelines, tags tend to be noisy and sparse. Especially new resources annotated by only a few users have often rather idiosyncratic tags that do not reflect a common perspective useful for search. In this paper we introduce an approach based on Latent Dirichlet Allocation (LDA) for recommending tags of resources in order to improve search. Resources annotated by many users and thus equipped with a fairly stable and complete tag set are used to elicit latent topics to which new resources with only a few tags are mapped. Based on this, other tags belonging to a topic can be recommended for the new resource. Our evaluation shows that the approach achieves significantly better precision and recall than the use of association rules, suggested in previous work, and also recommends more specific tags. Moreover, extending resources with these recommended tags significantly improves search for new resources.
TagiCoFi: tag informed collaborative filtering BIBAFull-Text 69-76
  Yi Zhen; Wu-Jun Li; Dit-Yan Yeung
Besides the rating information, an increasing number of modern recommender systems also allow the users to add personalized tags to the items. Such tagging information may provide very useful information for item recommendation, because the users' interests in items can be implicitly reflected by the tags that they often use. Although some content-based recommender systems have made preliminary attempts recently to utilize tagging information to improve the recommendation performance, few recommender systems based on collaborative filtering (CF) have employed tagging information to help the item recommendation procedure. In this paper, we propose a novel framework, called tag informed collaborative filtering (TagiCoFi), to seamlessly integrate tagging information into the CF procedure. Experimental results demonstrate that TagiCoFi outperforms its counterpart which discards the tagging information even when it is available, and achieves state-of-the-art performance.


Search shortcuts: a new approach to the recommendation of queries BIBAFull-Text 77-84
  Ranieri Baraglia; Fidel Cacheda; Victor Carneiro; Diego Fernandez; Vreixo Formoso; Raffaele Perego; Fabrizio Silvestri
The recommendation of queries, known as query suggestion, is a common practice on major Web Search Engines. It aims to help users to find the information they are looking for, and is usually based on the knowledge learned from past interactions with the search engine. In this paper we propose a new model for query suggestion, the Search Shortcut Problem, that consists in recommending "successful" queries that allowed other users to satisfy, in the past, similar information needs. This new model has several advantages with respect to traditional query suggestion approaches. First, it allows a straightforward evaluation of algorithms from available query log data. Moreover, it simplifies the application of several recommendation techniques from other domains. Particularly, in this work we applied Collaborative Filtering to this problem, and evaluated the interesting results achieved on large query logs from AOL and Microsoft. Different techniques for analyzing and extracting information from query logs, as well as new metrics and techniques for measuring the effectiveness of recommendations are proposed and evaluated. The results obtained clearly show the importance of several of our contributions, and open an interesting field for future research.
Increasing engagement through early recommender intervention BIBAFull-Text 85-92
  Jill Freyne; Michal Jacovi; Ido Guy; Werner Geyer
Social network sites rely on the contributions of their members to create a lively and enjoyable space. Recent research has focused on using personalization and recommender technologies to encourage participation of existing members. In this work we present an early-intervention approach to encouraging participation and engagement, which makes recommendations to new users during their sign-up process. Our recommender system exploits external social media to produce people and profile entry recommendations for new users. We present results of a live user study, showing that users who received recommendations at sign-up created more social connections, contributed more content, and were on the whole more engaged with the system, contributing more without prompt and returning more often. We further show that recommendations for multiple content types yield significantly better results, in terms of user contribution and consumption; and that recommendations of more active users yield a higher return rate.
Recommending new movies: even a few ratings are more valuable than metadata BIBAFull-Text 93-100
  István Pilászy; Domonkos Tikk
The Netflix Prize (NP) competition gave much attention to collaborative filtering (CF) approaches. Matrix factorization (MF) based CF approaches assign low dimensional feature vectors to users and items. We link CF and content-based filtering (CBF) by finding a linear transformation that transforms user or item descriptions so that they are as close as possible to the feature vectors generated by MF for CF. We propose methods for explicit feedback that are able to handle 140,000 features when feature vectors are very sparse. With movie metadata collected for the NP movies we show that the prediction performance of the methods is comparable to that of CF, and can be used to predict user preferences on new movies. We also investigate the value of movie metadata compared to movie ratings in regards of predictive power. We compare our solely CBF approach with a simple baseline rating-based predictor. We show that even 10 ratings of a new movie are more valuable than its metadata for predicting user ratings.
Regret-based optimal recommendation sets in conversational recommender systems BIBAFull-Text 101-108
  Paolo Viappiani; Craig Boutilier
Current conversational recommender systems are unable to offer guarantees on the quality of their recommendations due to a lack of principled user utility models. We develop an approach to recommender systems that incorporates an explicit utility model into the recommendation process in a decision-theoretically sound fashion. The system maintains explicit constraints on user utility based on preferences revealed by the user's actions. We investigate a new decision criterion, setwise minimax regret (SMR), for constructing optimal recommendation sets: we develop algorithms for computing SMR, and prove that SMR determines choice sets for queries that are myopically optimal. This provides a natural basis for generating compound critiques in conversational recommender systems. Our simulation results suggest that this utility-theoretically sound approach to user modeling allows much more effective navigation of a product space than traditional approaches based on, for example, heuristic utility models and product similarity measures.

Algorithms II

Stacking recommendation engines with additional meta-features BIBAFull-Text 109-116
  Xinlong Bao; Lawrence Bergman; Rich Thompson
In this paper, we apply stacking, an ensemble learning method, to the problem of building hybrid recommendation systems. We also introduce the novel idea of using runtime metrics which represent properties of the input users/items as additional meta-features, allowing us to combine component recommendation engines at runtime based on user/item characteristics. In our system, component engines are level-1 predictors, and a level-2 predictor is learned to generate the final prediction of the hybrid system. The input features of the level-2 predictor are predictions from component engines and the runtime metrics. Experimental results show that our system outperforms each single component engine as well as a static hybrid system. Our method has the additional advantage of removing restrictions on component engines that can be employed; any engine applicable to the target recommendation task can be easily plugged into the system.
A unified approach to building hybrid recommender systems BIBAFull-Text 117-124
  Asela Gunawardana; Christopher Meek
Content-based recommendation systems can provide recommendations for "cold-start" items for which little or no training data is available, but typically have lower accuracy than collaborative filtering systems. Conversely, collaborative filtering techniques often provide accurate recommendations, but fail on cold start items. Hybrid schemes attempt to combine these different kinds of information to yield better recommendations across the board.
   We describe unified Boltzmann machines, which are probabilistic models that combine collaborative and content information in a coherent manner. They encode collaborative and content information as features, and then learn weights that reflect how well each feature predicts user actions. In doing so, information of different types is automatically weighted, without the need for careful engineering of features or for post-hoc hybridization of distinct recommender systems.
   We present empirical results in the movie and shopping domains showing that unified Boltzmann machines can be used to combine content and collaborative information to yield results that are competitive with collaborative techniques in recommending items that have been seen before, and also effective at recommending cold-start items.
Exploiting user similarity based on rated-item pools for improved user-based collaborative filtering BIBAFull-Text 125-132
  Yue Shi; Martha Larson; Alan Hanjalic
An approach to user-based collaborative filtering is proposed that refines prediction of item ratings that is based on global user similarity by incorporating information derived from a more detailed user comparison made on the basis of Rated Item Pools (RIPs). The preference spectrum defined by items that a user has rated, and ranging from best-liked to most disliked items, is divided into item sets, or RIPs, which supply the basis for a fine-grained calculation of similarity between users. The RIP-based approach makes it possible for the model to take advantage of user tastes that are matched at one end of the spectrum, e.g., two users agree on favorites, without requiring complete correspondence of item ratings between user profiles. The approach improves rating prediction, as compared to a baseline that uses the global user similarity alone. It does not unduly inflate computational complexity or rely on external resources, common shortcomings of competing rating prediction methods. Cases in which the nearest neighbors are relatively dissimilar, known to be challenging for user-based collaborative filtering, demonstrate particularly substantial improvement. Performance is shown to be stable across the choice of neighborhood size, number of pools and relative pool size.
Assessment of conversation co-mentions as a resource for software module recommendation BIBAFull-Text 133-140
  Daniel Xiaodan Zhou; Paul Resnick
Conversation double pivots recommend target items related to a source item, based on co-mentions of source and target items in online forums. We deployed several variants on the drupal.org site that supports the Drupal open source community, and assessed them through clickthrough rates. A similarity metric based on correlation of mentions rather than mere co-occurrence reduced the problem of over-recommending the most popular modules, but additional corrections for recency and uniqueness of mentions were not helpful. Detection of more module mentions in conversations dramatically improved the quality of recommendations, even though the detection algorithm then had more false positives. Recommendations based on conversation co-mention were more effective than those based on co-installation, because co-installation data only led to recommendations of complementary modules and not substitutes. Recommendations based on co-mention were more effective than those based on text similarity matching for navigating from the most popular modules, but less effective than text matching for less popular modules.

Privacy and security

Effective diverse and obfuscated attacks on model-based recommender systems BIBAFull-Text 141-148
  Zunping Cheng; Neil Hurley
Robustness analysis research has shown that conventional memory-based recommender systems are very susceptible to malicious profile-injection attacks. A number of attack models have been proposed and studied and recent work has suggested that model-based collaborative filtering (CF) algorithms have greater robustness against these attacks. Moreover, to combat such attacks, several attack detection algorithms have been proposed. One that has shown high detection accuracy is based on using principal component analysis (PCA) to cluster attack profiles on the basis that such profiles are highly correlated. In this paper, we argue that the robustness observed in model-based algorithms is due to the fact that the proposed attacks have not targeted the specific vulnerabilities of these algorithms. We discuss how an effective attack targeting model-based algorithms that employ profile clustering can be designed. It transpires that the attack profiles employed in this attack, exhibit low rather than high pair-wise similarities and can easily be obfuscated to avoid PCA-based detection, while remaining effective.
Statistical attack detection BIBAFull-Text 149-156
  Neil Hurley; Zunping Cheng; Mi Zhang
It has been shown in recent years that effective profile injection or shilling attacks can be mounted on standard recommendation algorithms. These attacks consist of the insertion of bogus user profiles into the system database in order to manipulate the recommendation output, for example to promote or demote the predicted ratings for a particular product. A number of attack models have been proposed and some detection strategies to identify these attacks have been empirically evaluated. In this paper we show that the standard attack models can be readily detected using statistical detection techniques. We argue that insufficient consideration of the effectiveness of attacks under a constraint of statistical invariance has been taken in past research. In fact, it is possible to create effective attacks that are undetectable using the detection strategies proposed to date, including the PCA-based clustering strategy which has shown excellent performance against standard attacks. Nevertheless, these more advanced attacks can also be detected with careful design of a statistical detector. The question posed for future research is whether attack models that produce effective attack profiles that are statistically identical to genuine profiles are really possible.
Preserving privacy in collaborative filtering through distributed aggregation of offline profiles BIBAFull-Text 157-164
  Reza Shokri; Pedram Pedarsani; George Theodorakopoulos; Jean-Pierre Hubaux
In recommender systems, usually, a central server needs to have access to users' profiles in order to generate useful recommendations. Having this access, however, undermines the users' privacy.
   The more information is revealed to the server on the user-item relations, the lower the users' privacy is. Yet, hiding part of the profiles to increase the privacy comes at the cost of recommendation accuracy or difficulty of implementing the method. In this paper, we propose a distributed mechanism for users to augment their profiles in a way that obfuscates the user-item connection to an untrusted server, with minimum loss on the accuracy of the recommender system. We rely on the central server to generate the recommendations. However, each user stores his profile offline, modifies it by partly merging it with the profile of similar users through direct contact with them, and only then periodically uploads his profile to the server. We propose a metric to measure privacy at the system level, using graph matching concepts. Applying our method to the Netflix prize dataset, we show the effectiveness of the algorithm in solving the tradeoff between privacy and accuracy in recommender systems in an applicable way.
Manipulation-resistant collaborative filtering systems BIBAFull-Text 165-172
  Benjamin Van Roy; Xiang Yan
A collaborative filtering system recommends to users products that similar users like. Collaborative filtering systems influence purchase decisions, and hence have become targets of manipulation by unscrupulous vendors. We provide theoretical and empirical results demonstrating that while common nearest neighbor algorithms, which are widely used in commercial systems, can be highly susceptible to manipulation, a class of collaborative filtering algorithms which we refer to as linear is relatively robust. These results provide guidance for the design of future collaborative filtering systems.

Trust and evaluation

Rate it again: increasing recommendation accuracy by user re-rating BIBAFull-Text 173-180
  Xavier Amatriain; Josep M. Pujol; Nava Tintarev; Nuria Oliver
A common approach to designing Recommender Systems (RS) consists of asking users to explicitly rate items in order to collect feedback about their preferences. However, users have been shown to be inconsistent and to introduce a non-negligible amount of natural noise in their ratings that affects the accuracy of the predictions. In this paper, we present a novel approach to improve RS accuracy by reducing the natural noise in the input data via a preprocessing step. In order to quantitatively understand the impact of natural noise, we first analyze the response of common recommendation algorithms to this noise. Next, we propose a novel algorithm to denoise existing datasets by means of re-rating: i.e. by asking users to rate previously rated items again. This denoising step yields very significant accuracy improvements. However, re-rating all items in the original dataset is unpractical. Therefore, we study the accuracy gains obtained when re-rating only some of the ratings.In particular, we propose two partial denoising strategies: data and user-dependent denoising. Finally, we compare the value of adding a rating of an unseen item vs. re-rating an item. We conclude with a proposal for RS to improve the quality of their user data and hence their accuracy: asking users to re-rate items might, in some circumstances, be more beneficial than asking users to rate unseen items.
Using a trust network to improve top-N recommendation BIBAFull-Text 181-188
  Mohsen Jamali; Martin Ester
Top-N item recommendation is one of the important tasks of recommenders. Collaborative filtering is the most popular approach to building recommender systems which can predict ratings for a given user and item. Collaborative filtering can be extended for top-N recommendation, but this approach does not work accurately for cold start users that have rated only a very small number of items. In this paper we propose novel methods exploiting a trust network to improve the quality of top-N recommendation. The first method performs a random walk on the trust network, considering the similarity of users in its termination condition. The second method combines the collaborative filtering and trust-based approach. Our experimental evaluation on the Epinions dataset demonstrates that approaches using a trust network clearly outperform the collaborative filtering approach in terms of recall, in particular for cold start users.
Learning to recommend with trust and distrust relationships BIBAFull-Text 189-196
  Hao Ma; Michael R. Lyu; Irwin King
With the exponential growth of Web contents, Recommender System has become indispensable for discovering new information that might interest Web users. Despite their success in the industry, traditional recommender systems suffer from several problems. First, the sparseness of the user-item matrix seriously affects the recommendation quality. Second, traditional recommender systems ignore the connections among users, which loses the opportunity to provide more accurate and personalized recommendations. In this paper, aiming at providing more realistic and accurate recommendations, we propose a factor analysis-based optimization framework to incorporate the user trust and distrust relationships into the recommender systems. The contributions of this paper are three-fold: (1) We elaborate how user distrust information can benefit the recommender systems. (2) In terms of the trust relations, distinct from previous trust-aware recommender systems which are based on some heuristics, we systematically interpret how to constrain the objective function with trust regularization. (3) The experimental results show that the distrust relations among users are as important as the trust relations. The complexity analysis shows our method scales linearly with the number of observations, while the empirical analysis on a large Epinions dataset proves that our approaches perform better than the state-of-the-art approaches.
Personalised and dynamic trust in social networks BIBAFull-Text 197-204
  Frank E. Walter; Stefano Battiston; Frank Schweitzer
We propose a novel trust metric for social networks which is suitable for application to recommender systems. It is personalised and dynamic, and allows to compute the indirect trust between two agents which are not neighbours based on the direct trust between agents that are neighbours. In analogy to some personalised versions of PageRank, this metric makes use of the concept of feedback centrality and overcomes some of the limitations of other trust metrics. In particular, it does not neglect cycles and other patterns characterising social networks, as some other algorithms do. In order to apply the metric to recommender systems, we propose a way to make trust dynamic over time. We show by means of analytical approximations and computer simulations that the metric has the desired properties. Finally, we carry out an empirical validation on a dataset crawled from an Internet community and compare the performance of a recommender system using our metric to one using collaborative filtering.

Short papers

A case study on the effectiveness of recommendations in the mobile internet BIBAFull-Text 205-208
  Dietmar Jannach; Kolja Hegelich
This paper summarizes the initial findings of an experimental evaluation of how recommender systems affect the buying behavior of online customers. The study was conducted in the context of a large-scale, commercial Mobile Internet platform, from which end users can download games to their mobile phones. Item recommendations were presented to platform visitors in different navigational situations; the recommendation lists were either determined with the help of different recommendation algorithms or based on nonpersonalized ranking techniques. The study is based on a sample of more than 155,000 different customers who visited the portal during a four week evaluation period. The analysis revealed that the use of personalized recommendations instead of non-personalized ones leads to a significant increase in viewed and sold items in different navigational situations and to an overall sales increase.
A partial-order based active cache for recommender systems BIBAFull-Text 209-212
  Umar Qasim; Vincent Oria; Yi-Fang Brook Wu; Michael E. Houle; M. Tamer Özsu
Recommender systems aim to substantially reduce information overload by suggesting lists of similar items that users may find interesting.Caching has been a useful technique for reducing stress on limited resources and improving response time. In this paper, we propose an 'active caching' technique for recommender systems based on a partial order approach that not only benefits from popularity and temporal locality, but also exploits spatial locality. This approach allows the processing of answers to neighboring non-cached queries in addition to the reporting of cached query results. Test results for several data sets and recommendation techniques show substantial improvement in the cache hit ratio and computational costs, while achieving reasonable recall rates.
A recommender system for dynamically evolving online forums BIBAFull-Text 213-216
  Carlos Castro-Herrera; Jane Cleland-Huang; Bamshad Mobasher
Recommender systems can be used in online forums to recommend discussion topics to users; however as these forums are characterized by a constant influx of new users and new posts, it is important to consider the performance of the recommender system under a scenario in which the internal composition of the items to be recommended, i.e., discussion threads, and the user preferences are constantly changing. In this paper we describe and evaluate a forum recommender designed to handle the challenges of dynamically evolving internet forums used to gather and discuss feature requests for various software products. In particular, we empirically show that two proposed enhancements to the representations of user profiles will result in improved recommendation effectiveness in dynamic environments.
A semantic framework for personalized ad recommendation based on advanced textual analysis BIBAFull-Text 217-220
  Dorothea Tsatsou; Fotis Menemenis; Ioannis Kompatsiaris; Paul C. Davis
In this paper we present a hybrid recommendation system that combines ontological knowledge with content-extracted linguistic information, derived from pre-trained lexical graphs, in order to produce high quality, personalized recommendations. In the described approach, such recommendations are exemplified in an advertising scenario. We propose a distributed system architecture that uses semantic knowledge, based on terminologically enriched domain ontologies, to learn ontological user profiles and consequently infer recommendations through fuzzy semantic reasoning. A real world user study demonstrates the improvements attained in providing user-relevant recommendations with the aid of semantic profiles.
Acceptance issues of personality-based recommender systems BIBAFull-Text 221-224
  Rong Hu; Pearl Pu
To understand users' acceptance of the emerging trend of personality-based recommenders (PBR), we evaluated an existing PBR using the technology acceptance model (TAM). We also compare it with a baseline rating-based recommender in a within-subject user study. Our results show that while the personality-based recommender is perceived to be only slightly more accurate than the rating-based one, it is much easier to use. The side-by-side comparison also reveals that users significantly favor the personality-based recommender and have a significantly higher intention to use such a system again. Therefore, we believe that if users accepted rating-based recommenders, they are most likely to accept personality-based recommenders and personality-based recommenders have a high likelihood to be widely adopted despite the fact that rating-based recommenders are now the industry norm. We further point out some preliminary guidelines on how to design personality-based recommender systems.
Adaptive tradeoff explanations in conversational recommenders BIBAFull-Text 225-228
  Li Chen
The completeness and certainty of a user's preferences may vary during her preference construction process in a conversational recommender. In order to more effectively support users to uncover their hidden criteria and/or solve preference conflicts, we propose to generate adaptive tradeoff explanations in organization-based recommender interfaces, to be conditional on the user's contextual needs. An experiment shows the adaptive element's higher potential to improve recommendation efficiency, relative to methods without this feature.
An incentive-based architecture for social recommendations BIBAFull-Text 229-232
  Rajat Bhattacharjee; Ashish Goel; Konstantinos Kollias
We present an incentive-based architecture for providing recommendations in a social network. We maintain a distinct reputation system for each individual and we rely on users to identify appropriate correlations and rate the items using a system-provided recommendation language. The key idea is to design an incentive structure and a ranking system such that any inaccuracy in the recommendations implies the existence of a profitable arbitrage opportunity, hence making the system resistant to malicious spam and presentation bias. We also show that, under mild assumptions, our architecture provides users with incentive to minimize the Kullback-Leibler divergence between the ratings and the actual item qualities, quickly driving the system to an equilibrium state with accurate recommendations.
Analysis of cold-start recommendations in IPTV systems BIBAFull-Text 233-236
  Paolo Cremonesi; Roberto Turrin
In this paper we evaluate the performance of different collaborative algorithms in cold-start situations, where the initial lack of ratings may affect the quality of the algorithms. The evaluation has been performed on the pay-per-view datasets collected by two IP-television providers over a period of several months. The analysis shows that item-based algorithms perform better with respect to SVD-based ones in the early stage of the cold-start problem. Moreover, the accuracy of SVD-based algorithms, when using few latent factors, decreases with the time-evolution of the dataset. On the contrary, the same algorithms used with a large-enough number of latent features increase their accuracy with time and may outperform the item-based algorithms.
Collaborative filtering for social tagging systems: an experiment with CiteULike BIBAFull-Text 237-240
  Denis Parra; Peter Brusilovsky
Collaborative tagging systems pose new challenges to the developers of recommender systems. As observed by recent research, traditional implementations of classic recommender approaches, such as collaborative filtering, are not working well in this new context. To address these challenges, a number of research groups worldwide work on adapting these approaches to the specific nature of collaborative tagging systems. In joining this stream of research, we have developed and compared three variants of user-based collaborative filtering algorithms to provide recommendations of articles on CiteULike. The first approach, Classic Collaborative filtering (CCF) uses Pearson correlation to calculate similarity between users and a classic adjusted ratings formula to rank the recommendations. The second approach, Neighbor-weighted Collaborative Filtering, takes into account the number of raters in the ranking formula of the recommendations. The third approach explores an innovative way to form the user neighborhood based on a modified version of the Okapi BM25 model over users' tags. Our results suggest that both alterations of CCF are beneficial. Incorporating the number of raters into the algorithms leads to an improvement of precision, while tag-based BM25 can be considered as an alternative to Pearson correlation to calculate the similarity between users and their neighbors.
Conceptual recommender system for CiteSeerX BIBAFull-Text 241-244
  Ajith Kodakateri Pudhiyaveetil; Susan Gauch; Hiep Luong; Josh Eno
Short search engine queries do not provide contextual information, making it difficult for traditional search engines to understand what users are really requesting. One approach to this problem is to use recommender systems that identify user interests through various methods in order to provide information specific to the user's needs. However, many current recommender systems use a collaborative model based on a network of users to provide the recommendations, leading to problems in environments where network relationships are sparse or unknown. Content-based recommenders can avoid the sparsity problem but they may be inefficient for large document collections. In this paper, we propose a concept-based recommender system that recommends papers to general users of the CiteSeerX digital library of Computer Science research publications. We also represent a novel way of classifying documents and creating user profiles based on the ACM (Association for Computer Machinery) classification tree. Based on these user profiles which are built using past click histories, relevant papers in the domain are recommended to users. Experiments with a set of users on the CiteSeerX database show that our concept-based method provides accurate recommendations even with limited user profile histories.
Context-based splitting of item ratings in collaborative filtering BIBAFull-Text 245-248
  Linas Baltrunas; Francesco Ricci
Collaborative Filtering (CF) recommendations are computed by leveraging a historical data set of users' ratings for items. It assumes that the users' previously recorded ratings can help in predicting future ratings. This has been validated extensively, but in some domains item ratings can be influenced by contextual conditions, such as the time or the goal of the item consumption. This type of information is not exploited by standard CF models. This paper introduces and analyzes a novel pre-filtering technique for context-aware CF called item splitting. In this approach, the ratings of certain items are split, according to the value of an item-dependent contextual condition. Each split item generates two fictitious items that are used in the prediction algorithm instead of the original one. We evaluated this approach on real world and semi-synthetic data sets using matrix-factorization and nearest neighbor CF algorithms. We show that item splitting can be beneficial and its performance depends on the item selection method and on the influence of the contextual variables on the item ratings.
Critiquing recommenders for public taste products BIBAFull-Text 249-252
  Pearl Pu; Maoan Zhou; Sylvain Castagnos
Critiquing-based recommenders do not require users to state all of their preferences upfront or rate a set of previously experienced products. Compared to other types of recommenders, they require relatively little user effort, especially initially, despite potential accuracy problems. On the other hand, they rely on a set of critiques to elicit users feedback in order to improve accuracy. Thus the better the critiques are, the more accurately and efficiently the system becomes in generating its recommendations. This method has been successfully applied to high-involvement products. However, it was never tested on public taste products such as music, films, perfumes, fashion goods or wine. Indeed our initial trial adapting traditional critiquing methods to this new domain led to unsatisfactory results. This has motivated us to develop a novel approach named "editorial picked critiques" (EPC) that accounts for users' needs for popularity information, editorial suggestions, as well as their needs for personalization and diversity. Through an empirical study, we demonstrate that EPC presents a viable recommender approach and is superior on several dimensions to critiques generated by data mining methods.
Donation dashboard: a recommender system for donation portfolios BIBAFull-Text 253-256
  Tavi Nathanson; Ephrat Bitton; Ken Goldberg
In this paper we present Donation Dashboard, a system that recommends non-profit organizations to users in the form of a portfolio of donation amounts. Recommendations are made using our Eigentaste 2.0 constant-time collaborative filtering algorithm in combination with a new method for generating a weighted portfolio of recommendations. The key challenge is to generate a customized portfolio that does not necessarily exclude items already rated by the user. Under our method, the weights for items in the portfolio that have not yet been rated by the user are normalized factors of their predicted ratings, and the weights for items previously rated by the user are normalized factors of the actual ratings. Donation Dashboard 1.0 launched in April 2008, and as of May 8 2009 we have collected over 59,000 ratings of 70 nonprofit organizations from over 3,800 users.
   In this working paper we describe our experience developing Donation Dashboard, including the design of the system and our new method for portfolio generation. We use Normalized Mean Absolute Error (NMAE) to measure the accuracy of Eigentaste using our dataset of non-profit organization ratings and we compare that with the global mean algorithm. We analyze the data collected since the launch of the site, and we have made our dataset available to the public. Donation Dashboard and the Donation Dashboard dataset are accessible at: http://dd.berkeley.edu and http://dd.berkeley.edu/dataset
DynamicTV: a culture-aware recommender BIBAFull-Text 257-260
  Fabrizio Antonelli; Gianluca Francini; Marina Geymonat; Skjalg Lepsøy
A geographically homogeneous group of citizens shares much common knowledge, characteristics of their culture and history. This knowledge is captured for the use in an item-based recommender system that uses textual information, by introducing bias corpora: newspaper articles that represent the shared knowledge. We present a technique for incorporating and quickly replacing bias corpora in a case study of recommendation of TV contents on our IPTV platform. With this recommender, users watched more items and expressed satisfaction with the service.
Ensemble methods for improving the performance of neighborhood-based collaborative filtering BIBAFull-Text 261-264
  Alon Schclar; Alexander Tsikinovsky; Lior Rokach; Amnon Meisels; Liat Antwarg
Recommender systems provide consumers with ratings of items. These ratings are based on a set of ratings that were obtained from a wide scope of users. Predicting the ratings can be formulated as a regression problem. Ensemble regression methods are effective tools that improve the results of simple regression algorithms by iteratively applying the simple algorithm to a diverse set of inputs. The present paper describes a simple and effective ensemble regressor for the prediction of missing ratings in recommender systems. The ensemble method is an adaptation of the AdaBoost regression algorithm for recommendation tasks. In all iterations, interpolation weights for all nearest neighbors are simultaneously derived by minimizing the root mean squared error. From iteration to iteration instances that are hard to predict are reinforced by manipulating their weights in the goal function that needs to be minimized. The experimental evaluation demonstrates that the ensemble methodology significantly improves the predictive performance of single neighborhood-based collaborative filtering.
Experimental comparison of pre- vs. post-filtering approaches in context-aware recommender systems BIBAFull-Text 265-268
  Umberto Panniello; Alexander Tuzhilin; Michele Gorgoglione; Cosimo Palmisano; Anto Pedone
Recently, methods for generating context-aware recommendations were classified into the pre-filtering, post-filtering and contextual modeling approaches. Although some of these methods have been studied independently, no prior research compared the performance of these methods to determine which of them is better than the others. This paper focuses on comparing the pre-filtering and the post-filtering approaches and identifying which method dominates the other and under which circumstances. Since there are no clear winners in this comparison, we propose an alternative more effective method of selecting the winners in the pre- vs. the post-filtering comparison. This strategy provides analysts and companies with a practical suggestion on how to pick a good pre- or post-filtering approach in an effective manner to improve performance of a context-aware recommender system.
FeedbackTrust: using feedback effects in trust-based recommendation systems BIBAFull-Text 269-272
  Samaneh Moghaddam; Mohsen Jamali; Martin Ester; Jafar Habibi
With the advent of online social networks, the trust-based approach to recommendation has emerged which exploits the trust network among users and makes recommendations based on the ratings of trusted users in the network. In this paper, we introduce a two dimensional trust model which dynamically gets updated based on users's feedbacks, in contrast to static trust values in current trust models. Explorability measures the extent to which a user can rely on recommendations returned by the social network of a trusted user. Dependability represents the extent to which a user's own ratings can be trusted by users trusting him directly and indirectly. We propose a method to learn the values of explorability and dependability from raw trust data and feedback expressed by users on the recommendations they receive. Positive feedback will increase the trust and negative feedback will decrease the trust among users. We performed an evaluation on the Epinions dataset, demonstrating that exploiting user feedback results in lower prediction error compared to existing trust-based and collaborative filtering approaches.
FriendSensing: recommending friends using mobile phones BIBAFull-Text 273-276
  Daniele Quercia; Licia Capra
We propose FriendSensing, a framework that automatically suggests friends to mobile social-networking users. Using short-range technologies (e.g., Bluetooth) on her mobile phone, a social-networking user "senses" and keeps track of other phones in her proximity. FriendSensing processes proximity records using a variety of algorithms that are based on social network theories of geographical proximity and of link prediction. It then returns a personalized and automatically generated list of people the user may know. We evaluate the extent to which FriendSensing helps users find people they know against real mobility and social network data.
Generating comparative descriptions of places of interest in the tourism domain BIBAFull-Text 277-280
  Berardina De Carolis; Nicole Novielli; Vito Leonardo Plantamura; Enrica Gentile
When visiting cities as tourists, most of the times people do not make very detailed plans and, when choosing where to go and what to seem they tend to select the area with the major number of interesting facilities. Therefore, it would be useful to support the user choice with contextual information presentation, information clustering and comparative explanations of places of potential interest in a given area. In this paper we illustrate how MyMap, a mobile recommender system in the Tourism domain, generates comparative descriptions to support users in making decisions about what to see, among relevant objects of interest.
Generating transparent, steerable recommendations from textual descriptions of items BIBAFull-Text 281-284
  Stephen J. Green; Paul Lamere; Jeffrey Alexander; François Maillet; Susanna Kirk; Jessica Holt; Jackie Bourque; Xiao-Wen Mak
We propose a recommendation technique that works by collecting text descriptions of items and using this textual aura to compute the similarity between items using techniques drawn from information retrieval. We show how this representation can be used to explain the similarities between items using terms from the textual aura and further how it can be used to steer the recommender. We describe a system that demonstrates these techniques and we'll detail some preliminary experiments aimed at evaluating the quality of the recommendations and the effectiveness of the explanations of item similarity.
Getting recommender systems to think outside the box BIBAFull-Text 285-288
  Zeinab Abbassi; Sihem Amer-Yahia; Laks V. S. Lakshmanan; Sergei Vassilvitskii; Cong Yu
We examine the case of over-specialization in recommender systems, which results from returning items that are too similar to those previously rated by the user. We propose Outside-The-Box (otb) recommendation, which takes some risk to help users make fresh discoveries, while maintaining high relevance. The proposed formalization relies on item regions and attempts to identify regions that are under-exposed to the user. We develop a recommendation algorithm which achieves a compromise between relevance and risk to find otb items. We evaluate this approach on the MovieLens data set and compare our otb recommendations against conventional recommendation strategies.
Harnessing the power of "favorites" lists for recommendation systems BIBAFull-Text 289-292
  Maryam Khezrzadeh; Alex Thomo; William W. Wadge
We propose a novel collaborative recommendation approach to take advantage of the information available in user-created lists. Our approach assumes associations among any two items appearing in a list together. We calculate sum of Bayesian ratings (SBR) of all lists containing an item pair as the strength of item-item associations in that pair. SBR takes into consideration not only the number of lists the items have co-appeared in, but also the quality of the lists. We collected a data set of user ratings for books along with Listmania lists on Amazon.com using Amazon Web Services (AWS). Our method shows superior performance to existing user-based and item-based collaborative filtering approaches according to the resulted MAE, coverage and F-measure.
How does high dimensionality affect collaborative filtering? BIBAFull-Text 293-296
  Alexandros Nanopoulos; A Milos Radovanovic; A Mirjana Ivanovic
A crucial operation in memory-based collaborative filtering (CF) is determining nearest neighbors (NNs) of users/items. This paper addresses two phenomena that emerge when CF algorithms perform NN search in high-dimensional spaces that are typical in CF applications. The first is similarity concentration and the second is the appearance of hubs (i.e. points which appear in $k$-NN lists of many other points). Through theoretical analysis and experimental evaluation we show that these phenomena are inherent properties of high-dimensional space, unrelated to other data properties like sparsity, and that they can impact CF algorithms by questioning the meaning and representativeness of discovered NNs. Moreover, we show that it is not easy to mitigate the phenomena using dimensionality reduction. Studying these phenomena aims to provide a better understanding of the limitations of memory-based CF and motivate the development of new algorithms that would overcome them.
iTag: a personalized blog tagger BIBAFull-Text 297-300
  Michael Hart; Rob Johnson; Amanda Stent
We present iTag, a personalized tag recommendation system for blogs. iTag improves on current tag recommendation systems in two ways. First, iTag has much higher precision and recall than previously proposed tagging algorithms. For example, iTag achieved over 60% precision and recall on a set of 1000 blog posts selected at random from a WordPress RSS feed in April 2009, whereas the previously developed TagAssist achieved less than 10% precision and recall on our data. Second, iTag performs just as well when trained on a single user's blog as when trained on a large corpus of blogs. Thus, iTag can be deployed as a global, non-personalized tag recommendation system, or as a personalized tag recommender. Our experiments and survey of tagging behavior suggest that bloggers use tags idiosyncratically, so personalized tagging is an important option.
Knowledge infusion into content-based recommender systems BIBAFull-Text 301-304
  Giovanni Semeraro; Pasquale Lops; Pierpaolo Basile; Marco de Gemmis
Content-based recommender systems try to recommend items similar to those a given user has liked in the past. The basic process consists of matching up the attributes of a user profile, in which preferences and interests are stored, with the attributes of a content object (item).
   Common-sense and domain-specific knowledge may be useful to give some meaning to the content of items, thus helping to generate more informative features than "plain" attributes.
   The process of learning user profiles could also benefit from the infusion of exogenous knowledge or open source knowledge, with respect to the classical use of endogenous knowledge (extracted from the items themselves).
   The main contribution of this paper is a proposal for knowledge infusion into content-based recommender systems, which suggests a novel view of this type of systems, mostly oriented to content interpretation by way of the infused knowledge.
   The idea is to provide the system with the "linguistic" and "cultural" background knowledge that hopefully allows a more accurate content analysis than classic approaches based on words. A set of knowledge sources is modeled to create a memory of linguistic competencies and of more specific world "facts", that can be exploited to reason about content as well as to support the user profiling and recommendation processes.
   The modeled knowledge sources include a dictionary, Wikipedia, and content generated by users (i.e. tags provided on items), while the core of the reasoning component is a spreading activation algorithm.
Learning to recommend helpful hotel reviews BIBAFull-Text 305-308
  Michael P. O'Mahony; Barry Smyth
User-generated reviews are a common and valuable source of product information, yet little attention has been paid as to how best to present them to end-users. In this paper, we describe a classification-based recommender system that is designed to recommend the most helpful reviews for a given product. We present a large-scale evaluation of our approach using TripAdvisor hotel reviews, and we show that our approach is capable of suggesting superior reviews compared to a number of alternative recommendation benchmarks.
Maximum margin matrix factorization for code recommendation BIBAFull-Text 309-312
  Markus Weimer; Alexandros Karatzoglou; Marcel Bruch
Code recommender systems ease the use and learning of software frameworks and libraries by recommending calls based on already present code. Typically, code recommender tools have been based on rather simple rule based systems while many of the recent advances in Recommender Systems and Collaborative Filtering have been largely focused on rating data. While many of these advances can be incorporated in the code recommendation setting this problem also brings considerable challenges of its own. In this paper, we extend state-of-the-art collaborative filtering technology, namely Maximum Margin Matrix Factorization (MMMF) to this interesting application domain and show how to deal with the challenges posed by this problem. To this end, we introduce two new loss functions to the MMMF model. While we focus on code recommendation in this paper, our contributions and the methodology we propose can be of use in almost any collaborative setting that can be represented as a binary interaction matrix. We evaluate the algorithm on real data drawn from the Eclipse Open Source Project. The results show a significant improvement over current rule-based approaches.
Measuring predictive capability in collaborative filtering BIBAFull-Text 313-316
  Luis M. de Campos; Juan M. Fernandez-Luna; Juan F. Huete; Miguel A. Rueda-Morales
This paper presents a new memory-based approach to Collaborative Filtering where the neighbors of the active user will be selected taking into account their predictive capability. Our hypothesis is that if a user was good at predicting the past ratings, then his/her predictions will be also helpful to recommend ratings in the future. The predictive capability of a user will be measured using two different criteria: The first one which is based on the likelihood of the active user's rating and the second one tries to minimize the error obtained using his/her predictions. We show our experimental results using standard data sets.
MoviExplain: a recommender system with explanations BIBAFull-Text 317-320
  Panagiotis Symeonidis; Alexandros Nanopoulos; Yannis Manolopoulos
Providing justification to a recommendation gives credibility to a recommender system. Some recommender systems (Amazon.com etc.) try to explain their recommendations, in an effort to regain customer acceptance and trust. But their explanations are poor, because they are based solely on rating data, ignoring the content data. Our prototype system MoviExplain is a movie recommender system that provides both accurate and justifiable recommendations.
On the limitations of browsing top-N recommender systems BIBAFull-Text 321-324
  Klaus Seyerlehner; Arthur Flexer; Gerhard Widmer
To exploit the enormous potential of niche products, modern information systems must support users in exploring digital libraries and online catalogs. A straight-forward way of doing so is to support browsing the available items, which is in general realized by presenting a user the top-N recommendations for each item. However, recent research indicates that most of the niche products reside in the so-called Long Tail, and simple collaborative filtering-based recommender systems alone do not allow to explore these niche products. In this paper we show that it is not only a popularity problem related to the collaborative filtering approach that makes a portion of the elements of a digital library inaccessible via browsing, but also a consequence of the top N-recommendation approach itself.
Personality aware recommendations to groups BIBAFull-Text 325-328
  Juan A. Recio-Garcia; Guillermo Jimenez-Diaz; Antonio A. Sanchez-Ruiz; Belen Diaz-Agudo
In this article we introduce a novel method of making recommendations to groups based on existing techniques of collaborative filtering and taking into account the group personality composition. We have tested our method in the movie recommendation domain and we have experimentally evaluated its behavior under heterogeneous groups according to the group personality composition.
Personalized recommendation based on the personal innovator degree BIBAFull-Text 329-332
  Noriaki Kawamae; Hitoshi Sakano; Takeshi Yamada
This paper proposes a novel Collaborative Filtering scheme; it focuses on the dynamics and precedence of user preference to recommend items that match the latest preference of the target user. In predicting which items this user will purchase in the near future, the proposed algorithm identifies purchase history logs of users who have similar preferences and a high degree of purchase precedence (i.e., purchasing the same items earlier) relative to the target user. We call this metric the Personal Innovator Degree (PID). Experiments using real online sales data show that the proposed method outperforms existing methods.
Predicting social-tags for cold start book recommendations BIBAFull-Text 333-336
  Sharon Givon; Victor Lavrenko
We demonstrate how user ratings can be accurately predicted from a set of tags assigned to a book on a social-networking site. Since a newly-published book is unlikely to have social-tags already assigned to it, we describe a probabilistic model for inferring the most probable tags from the text of the book. We evaluate the proposed approach on a newly-created corpus, involving 146 books and 1060 users. Our experiments demonstrate that the proposed approach is significantly better than a well-tuned collaborative filtering baseline for books with 10 or fewer ratings. We also show how predictions based on social-tags can be combined with the traditional collaborative-filtering methods to yield superior performance with any number of ratings.
Preference aggregation in group recommender systems for committee decision-making BIBAFull-Text 337-340
  Jacob P. Baskin; Shriram Krishnamurthi
We present a preference aggregation algorithm designed for situations in which a limited number of users each review a small subset of a large (but finite) set of candidates. This algorithm aggregates scores by using users' relative preferences to search for a Kemeny-optimal ordering of items, and then uses this ordering to identify good and bad items, as well as those that are the subject of reviewer conflict. The algorithm uses variable-neighborhood local search, allowing the efficient discovery of high-quality consensus orderings while remaining computationally feasible. It provides a significant increase in solution quality over existing systems. We discuss potential applications of this algorithm in group recommender systems for a variety of scenarios, including program committees and faculty searches.
Preference elicitation with subjective features BIBAFull-Text 341-344
  Craig Boutilier; Kevin Regan; Paolo Viappiani
Utility or preference elicitation is a critical component in many recommender and decision support systems. However, most frameworks for elicitation assume a predefined set of features (e.g., as derived from catalog descriptions) over which user preferences are expressed. Just as user preferences vary considerably, so too can the features over which they are most comfortable expressing these preferences. In this work, we consider preference elicitation in the presence of subjective or user-defined features. We treat the problem of learning a user's feature definition as one of concept learning, but whose goal is to learn only enough about the concept definition to enable a good decision to be made. This is complicated by the fact that user preferences are unknown. We describe computational procedures for identifying optimal alternatives w.r.t minimax regret in the presence of both utility and concept uncertainty; and develop several heuristic query strategies that focus on reduction of relevant concept and utility uncertainty. Computational experiments verify the efficacy of these strategies.
Putting recommendations on the map: visualizing clusters and relations BIBAFull-Text 345-348
  Emden Gansner; Yifan Hu; Stephen Kobourov; Chris Volinsky
For users, recommendations can sometimes seem odd or counterintuitive. Visualizing recommendations can remove some of this mystery, showing how a recommendation is grouped with other choices. A drawing can also lead a user's eye to other options. Traditional 2D-embeddings of points can be used to create a basic layout, but these methods, by themselves, do not illustrate clusters and neighborhoods very well. In this paper, we propose the use of geographic maps to enhance the definition of clusters and neighborhoods, and consider the effectiveness of this approach in visualizing similarities and recommendations arising from TV shows.
Rating aggregation in collaborative filtering systems BIBAFull-Text 349-352
  Florent Garcin; Boi Faltings; Radu Jurca; Nadine Joswig
Recommender systems based on user feedback rank items by aggregating users' ratings in order to select those that are ranked highest. Ratings are usually aggregated using a weighted arithmetic mean. However, the mean is quite sensitive to outliers and biases, and thus may not be the most informative aggregate. We compare the accuracy and robustness of three different aggregators: the mean, median and mode. The results show that the median may often be a better choice than the mean, and can significantly improve recommendation accuracy and robustness in collaborative filtering systems.
Recommendations with prerequisites BIBAFull-Text 353-356
  Aditya G. Parameswaran; Hector Garcia-Molina
We consider the problem of recommending the best set of k items when there is an inherent ordering between items, expressed as a set of prerequisites (e.g., the course `Real Analysis' is a prerequisite of `Complex Analysis'). Since this problem is NP-hard, we develop 3 approximate algorithms to solve this problem. We experimentally evaluate these algorithms on synthetic data.
Recommender systems for the conference paper assignment problem BIBAFull-Text 357-360
  Don Conry; Yehuda Koren; Naren Ramakrishnan
We present a recommender systems approach to conference paper assignment, i.e., the task of assigning paper submissions to reviewers. We address both the modeling of reviewer-paper preferences (which can be cast as a learning problem) and the optimization of reviewing assignments to satisfy global conference criteria (which can be viewed as constraint satisfaction). Due to the paucity of preference data per reviewer or per paper (relative to other recommender systems applications) we show how we can integrate multiple sources of information to learn reviewer-paper preference models. Our models are evaluated not just in terms of prediction accuracy but in terms of end-assignment quality. Using a linear programming-based assignment optimization, we show how our approach better explores the space of unsupplied assignments to maximize the overall affinities of papers assigned to reviewers. We demonstrate our results on real reviewer bidding data from the IEEE ICDM 2007 conference.
Recommenders' influence on buyers' decision process BIBAFull-Text 361-364
  Sylvain Castagnos; Nicolas Jones; Pearl Pu
Online stores offer an increasingly large set of products. Interactive decision aids are becoming indispensable tools assisting users as they search for an ideal product to purchase. For an e-commerce website, adopting the correct tools can affect its survival: effective product recommender tools are increasingly recognized by online stores as effective means to sell more products; on the other hand, sites that do not employ intelligent tools will not only see poor purchase volumes but also experience less traffic because consumers are more likely to return to a site employing recommender systems.
   This paper presents ongoing research in understanding the impact of various decision aids on users' interaction behaviors and their subjective perceptions of these aids. In the current experiment, we employed an eye tracker in an in-depth user study to understand the influence of recommenders on how users select items for the basket set. We collected more than 20,300 fixation data points in 3,648 areas of interest. Our studies show that while users still rely on product filtering tools, the use of recommenders is becoming more prominent in helping them construct the basket set and is monotonically increasing as time goes on.
Team recommendation in open innovation networks BIBAFull-Text 365-368
  Michele Brocco; Georg Groh
Open Innovation has become an important new paradigm for incorporating external knowledge and sources in the innovation process of organizations. Besides other discussed arguments the resulting large size of innovator networks suggests that algorithmic approaches for team recommendation may be needed in that scenario. The current work identifies the related difficulties and thoroughly investigates aspects entities for the problem of team recommendation. Based on that, we develop a meta model which allows to instantiate and integrate most of the vast number of the existing socio-/psychological models on optimal team composition. This meta model is necessary for operationalizing our intended team recommendation approach.
Testing and evaluating tag recommenders in a live system BIBAFull-Text 369-372
  Robert Jäschke; Folke Eisterlehner; Andreas Hotho; Gerd Stumme
The challenge to provide tag recommendations for collaborative tagging systems has attracted quite some attention of researchers lately. However, most research focused on the evaluation and development of appropriate methods rather than tackling the practical challenges of how to integrate recommendation methods into real tagging systems, record and evaluate their performance. In this paper we describe the tag recommendation framework we developed for our social bookmark and publication sharing system BibSonomy. With the intention to develop, test, and evaluate recommendation algorithms and supporting cooperation with researchers, we designed the framework to be easily extensible, open for a variety of methods, and usable independent from BibSonomy. Furthermore, this paper presents a first evaluation of two exemplarily deployed recommendation methods.
The 3A contextual ranking system: simultaneously recommending actors, assets, and group activities BIBAFull-Text 373-376
  Sandy El Helou; Christophe Salzmann; Stéphane Sire; Denis Gillet
In this paper, we propose a personalized and contextual ranking algorithm implemented on top of the 3A interaction model. The latter is a generic model intended for designing and describing social and collaborative learning platforms integrating Actors, Assets and group Activities (the 3 "A"). The target user's interactions with his/her environment are modeled in a heterogeneous graph. Then, the algorithm is applied to simultaneously rank actors, assets and group activities taking into account the target user and his/her context. As an illustrative application and a preliminary evaluation, we apply the algorithm on data related to the activities carried out in a European Research Project, especially the collaboration between its members through the joint production of deliverables in workpackages.
Uncovering functional dependencies in MDD-compiled product catalogues BIBAFull-Text 377-380
  Tarik Hadzic; Barry O'Sullivan
A functional dependency is a logical relationship amongst the attributes that define a table of data. Specifically, a functional dependency holds when the values of a subset of the attributes in a dataset determine the values of one or more other attributes. Uncovering such dependencies is utilized in many domains, such as database design. We demonstrate that it can also be utilized in a recommendation context when datasets represent product catalogues. State-of-the-art approaches to discovering functional dependencies require a tabular representation of the data. However, product catalogues can sometimes be defined implicitly, for example, as a set of solutions to a combinatorial problem. Such combinatorial catalogues can have a very large number of products, thus making standard approaches to uncovering functional dependencies inapplicable. In this paper we present the first approach to computing functional dependencies over compiled knowledge representations which can often be small even for huge catalogues. In particular, we develop efficient algorithms that operate over decision diagrams, which allow us to handle catalogues that are out of reach for current approaches. We apply our algorithms to tabular and combinatorial benchmarks and detect a number of properties that could be considered as anomalies in product catalogues.
Understanding the effect of adaptive preference elicitation methods on user satisfaction of a recommender system BIBAFull-Text 381-384
  Bart P. Knijnenburg; Martijn C. Willemsen
In a recommender system that suggests options based on user attribute weights, the method of preference elicitation (PE) employed by a recommender system can influence users' satisfaction with the system, as well as the perceived usefulness and the understandability of the system. Specifically, we hypothesize that users with different levels of domain knowledge prefer different types of PE. While domain experts reported higher satisfaction and perceived usefulness with attribute-based PE (i.e., indicating preference levels for the domain-related attributes), novices preferred case-based PE (i.e., indicating the preference for specific examples, from which attribute-preferences can then be implicitly calculated). The paper discusses the decision-theoretical principles that are believed to lead to this distinction, as well as an experiment that provides substantial evidence for the hypothesis. Consequently, we introduce the idea of adapting the method of PE to users' domain knowledge on the fly using click stream data.
Using twitter to recommend real-time topical news BIBAFull-Text 385-388
  Owen Phelan; Kevin McCarthy; Barry Smyth
Recommending news stories to users, based on their preferences, has long been a favourite domain for recommender systems research. In this paper, we describe a novel approach to news recommendation that harnesses real-time micro-blogging activity, from a service such as Twitter, as the basis for promoting news stories from a user's favourite RSS feeds. A preliminary evaluation is carried out on an implementation of this technique that shows promising results.
View-based recommender systems BIBAFull-Text 389-392
  Fabián P. P. Lousame; Eduardo Sánchez
Different recommender systems based on collaborative technology have been proposed that recommend new relevant products to users by exploring past user preference patterns. The most common approach generates recommendations based on user consumption patterns and on rating information gathered during each user-item interaction. In this paper we introduce a novel approach based on views of items, which are basically more complex models of the user-item interactions aimed at capturing the aspects on which users base their ratings. The resulting view-based approach recommends views to users instead of the traditional items. The proposed algorithms are tested on an artificial database, and the results show that modeling further interaction information improves the accuracy of predictions, provides a robust background to explain recommendations, exposes users to more specific recommendations and leads to better models of user preferences.

Doctoral symposium

A spatial model for collaborative filtering of comments in an online discussion forum BIBAFull-Text 393-396
  Ephrat Bitton
This extended abstract presents a new spatial model for collaboratively recommending "compelling" comments in an online discussion forum that promote consensus among a diverse group of users. In this application our goal is to promote comments that are rated highly by dissimilar users, which in some sense is a dual to traditional recommender problems. We propose a model for weighting and aggregating comment ratings that gives greater influence to positive ratings from users who tend to disagree with the commenter, and we compare it with various alternate methods. The model has the added benefit of being resistant to manipulation by false ratings and sybil attacks.
   We test the model on comments in Opinion Space, a new online discussion tool that allows users to visualize where they stand with respect to other users in terms of their opinions on a set of controversial propositions. Comments in the system are recommended visually, where more "compelling" comments are emphasized with larger sizes.
Enhancing diversity in Top-N recommendation BIBAFull-Text 397-400
  Mi Zhang
In recent years it has been argued that, besides the standard accuracy metrics, other characteristics of the recommendation algorithm ought to be taken into account when evaluating recommendation performance. One such characteristic is recommendation diversity and this topic is the focus of this research project. The overall goal of the project is to examine ways to improve the diversity of recommendations while maintaining high accuracy. During the course of my work to date I have addressed the question of how best to evaluate diversification strategies and have proposed a number of new diversity enhancement algorithms.
Applying relevant set correlation clustering to multi-criteria recommender systems BIBAFull-Text 401-404
  Nkechi J. Nnadi
This thesis investigates application of clustering to multi-criteria ratings as a method of improving the precision of top-N recommendations. With the advent of ecommerce sites that allow multi-criteria rating of items, there is an opportunity for recommender systems to use the additional information to gain a better understanding of user preference. This thesis proposes the use of the relevant set correlation model for a clustering-based collaborative filtering system. It is anticipated this novel system will handle large numbers of users and items without sacrificing the relevance of recommended items.
Matching information content with music BIBAFull-Text 405-408
  Marius Kaminskas
This work addresses a particular kind of cross domain personalization task consisting of selecting simultaneously two items in two different domains and recommending them together because they fit the user preferences and also they fit well together. In particular, it is shown that given a personalized recommendation for points of interests (POIs), the user's selection of POIs can be influenced by enriching their presentation with music tracks that are matching the user's profile and also the POIs. This paper presents the results of an online, live-users, experiment where we evaluated alternative approaches for matching POIs and music, based on tagging and text matching.
Personalized query relaxations and repairs in knowledge-based recommendation BIBAFull-Text 409-412
  Monika Schubert
Knowledge-based recommender systems are applications that support users in the process of retrieving items from a complex product assortment (e.g. computers, holiday packages, and financial services). Recommendations are determined on the basis of explicitly defined user requirements which can be interpreted as constraints to be fulfilled by the items stored in a product table. If no solution (item) can be found, existing knowledge-based recommenders propose non-personalized query relaxations and repair actions for the given set of customer requirements that support a recovery from the dead-end. This paper points out how the usability of knowledge-based recommender systems can be improved by introducing the concept of personalized query relaxations and repair actions.
Situation-aware trust management BIBAFull-Text 413-416
  Mozhgan Tavakolifard
We present a knowledge-intensive and model-based case-based reasoning framework that supports the truster for situation-aware trust management. The suggested method augments the typically sparse trust information by inferring the missing information from other situational conditions, and can better support situation-aware trust management. Our framework can be coupled with existing trust management models to make them situation-aware. It uses the underlying model of trust management to transfer trust information between situations. We validate the proposed framework for Subjective Logic trust management model and evaluate it by conducting experiments on a large real dataset.
Search the web x.0: mining and recommending web-mediated processes BIBAFull-Text 417-420
  Gabriele Tolomei
Nowadays, people have been increasingly interested in exploiting Web Search Engines (WSEs) not only for having access to simple Web pages, but mainly for accomplishing even complex activities, namely Web-mediated processes (or taskflows). Thus, users' information needs will become more complex, and Web search and recommender systems should change accordingly for dealing with this shift. We claim that such taskflows and their composing tasks are implicitly present in users' minds when they interact with a WSE to access the Web. Our first research challenge is thus to evaluate this belief by analyzing a very large, long-term log of queries submitted to a WSE, and associating meaningful semantic labels with the extracted tasks (i.e., clusters of related queries) and taskflows. This large knowledge base constitutes a good starting point for building a model of users' behaviors. The second research challenge is to devise a novel recommender system that goes beyond the simple query suggestion of modern WSEs. Our system has to exploit the knowledge base of Web-mediated processes and the learned model of users' behaviors, to generate complex insights and task-based suggestions to incoming users while they interact with a WSE.


Workshop on recommender systems and the social web BIBFull-Text 421-422
  Dietmar Jannach; Werner Geyer; Casey Dugan; Jill Freyne; Sarabjot Singh Anand; Bamshad Mobasher; Alfred Kobsa
RecSys'09 workshop 3: workshop on context-aware recommender systems (CARS-2009) BIBFull-Text 423-424
  Gediminas Adomavicius; Francesco Ricci


Tutorial on using social trust for recommender systems BIBAFull-Text 425-426
  Jennifer Golbeck
As the Web has shifted to an interactive environment where vast amounts of content is created by users, the question of whom to trust and what information to trust has become both more important and more difficult to answer. At the same time, social networks have become very popular with over a billion accounts shared across hundreds of networks. Social trust relationships, derived from social networks, are uniquely suited to speak to the quality of online information; recommender systems are designed to personalize, sort, aggregate, and highlight information. Merging social networks, trust, and recommender systems can improve the accuracy of recommendations and improve the user's experience. In this tutorial, we will cover the use of social trust in recommender systems. Topics including the computation of trust in social networks, integration of trust into recommender systems, and a discussion of when trust offers benefits and the challenges it presents.