Mouzhi Ge

Beyond accuracy: evaluating recommender systems by coverage and serendipity

When we evaluate the quality of recommender systems (RS), most approaches only focus on the predictive accuracy of these systems. Recent works suggest that beyond accuracy there is a variety of other metrics that should be considered when evaluating a RS. In this paper we focus on two crucial metrics in RS evaluation: coverage and serendipity. Based on a literature review, we first discuss both measurement methods as well as the trade-off between good coverage and serendipity. We then analyze the role of coverage and serendipity as indicators of recommendation quality, present novel ways of how they can be measured and discuss how to interpret the obtained measurements. Overall, we argue that our new ways of measuring these concepts reflect the quality impression perceived by the user in a better way than previous metrics thus leading to enhanced user satisfaction.

Recommender Systems in Computer Science and Information Systems – A Landscape of Research

The paper reviews and classifies recent research in recommender systems both in the field of Computer Science and Information Systems. The goal of this work is to identify existing trends, open issues and possible directions for future research. Our analysis is based on a review of 330 papers on recommender systems, which were published in high-impact conferences and journals during the past five years (2006-2011). We provide a state-of-the-art review on recommender systems, propose future research opportunities for recommender systems in both computer science and information system community, and indicate how the research avenues of both communities might partly converge.

Health-aware Food Recommender System

With the rapid changes in the food variety and lifestyles, many people are facing the problem of making healthier food decisions to reduce the risk of chronic diseases such as obesity and diabetes. To this end, our recommender system not only offers recipe recommendations that suit the users preference but is also able to take the users health into account. It is developed on a mobile platform by considering that our application may be directly used in the kitchen. This demo paper summarizes the complete human-computer interaction design, the implemented health-aware recommendation algorithm and preliminary user feedback.

Data and Information Quality Assessment in Information Manufacturing Systems

Organizations are more and more concerned about the increasing data and information quality issues in their information (manufacturing) systems. These issues have caused various organizational problems such as losing customers, missing opportunities and making incorrect decisions. Recognizing these issues, one of the crucial aspects for organizations to sustain business growth and competitive advantage is to be able to assess data and information quality. However limited research has been done to investigate data and information quality assessment in information manufacturing systems. This paper proposes a model to assess the quality of two major information sources in information manufacturing systems: data stored in database and information products delivered to users. The proposed model is applied to an information manufacturing system and an example database. The research findings have shown that the poor quality of data found in example databases is correlated to the quality of information products perceived by users.

Effects of information quality on inventory management

This paper investigated how different IQ categories affect the quality of inventory control decisions. The result has shown that intrinsic IQ and contextual IQ are positively related to inventory decision quality, indicating that intrinsic IQ and contextual IQ are the underlying concerns in improving inventory decision quality. It is also found that the effect of representational IQ on inventory decision quality is non-significant. However, under just-in-time inventory policy, representational IQ and contextual IQ jointly affect inventory decision quality. This result indicates that improving representational IQ may intensify the positive effects of contextual IQ.

RF-Rec: Fast and Accurate Computation of Recommendations Based on Rating Frequencies

The goal of recommender systems (RS) is to provide personalized recommendations of products or services to users facing the problem of information overload on the Web. The most popular approaches to retrieve the most relevant items for a user are collaborative filtering (CF) recommendation algorithms and in particular in recent years a number of sophisticated algorithms based, e.g., on matrix factorization or machine learning, have been proposed to improve the predictive accuracy of RS. In our recent work, we proposed a novel recommendation scheme called RF-Rec, which generates predictions simply by counting and combining the frequencies of the different rating values in the usual user-item rating matrix. The scheme has some key advantages when compared with more sophisticated techniques. It is trivial to implement, can generate predictions in constant time and has a high prediction coverage. In this paper we propose extensions to our method in order to further increase the predictive accuracy by introducing schemes to weight and parameterize the components of the predictor. An evaluation on three standard test data sets reveals that the accuracy of our new schemes is higher than traditional CF algorithms in particular on sparse data sets and on a par with a recent matrix factorization algorithm. At the same time, the key advantages of the basic scheme such as computational efficiency, scalability, simplicity and the support for incremental updates are still maintained.

Context Dependent Preference Acquisition with Personality-Based Active Learning in Mobile Recommender Systems

Nowadays, Recommender Systems (RSs) play a key role in many businesses. They provide consumers with relevant recommendations, e.g., Places of Interest (POIs) to a tourist, based on user preference data, mainly in the form of ratings for items. The accuracy of recommendations largely depends on the quality and quantity of the ratings (preferences) provided by the users. However, users often tend to rate no or only few items, causing low accuracy of the recommendation. Active Learning (AL) addresses this problem by actively selecting items to be presented to the user in order to acquire a larger number of high-quality ratings (preferences), and hence, improve the recommendation accuracy. In this paper, we propose a personalized active learning approach that leverages user’s personality data to get more and better in-context ratings. We have designed a novel human computer interaction and assessed our proposed approach in a live user study - which is not common in active learning research. The main result is that the system is able to collect better ratings and provide more relevant recommendations compared to a variant that is using a state of the art approach to preference acquisition.

Understanding Recommendations by Reading the Clouds

Current research has shown the important role of explanation facilities in recommender systems based on the observation that explanations can significantly influence the user-perceived quality of such a system. In this paper we present and evaluate explanation interfaces in the form of tag clouds, which are a frequently used visualization and interaction technique on the Web. We report the result of a user study in which we compare the performance of two new explanation methods based on personalized and non-personalized tag clouds with a previous explanation approach. Overall, the results show that explanations based on tag clouds are not only well-accepted by the users but can also help to improve the efficiency and effectiveness of the explanation process. Furthermore, we provide first insights on the value of personalizing explanations based on the recently-proposed concept of item-specific tag preferences.

Impact of Information Quality on Supply Chain Decisions

A number of studies suggest that making correct decisions depends on high-quality information; how information quality affects decision-making is still not fully understood. Following the multi-dimensional view of information quality, this paper investigates the effects of information accuracy, completeness, and consistency on decision-making. Results show that information accuracy and completeness affect decision quality significantly. Although the effect of information consistency on decision quality appears to be non-significant, consistency of information may intensify the contribution of accuracy, indicating that information accuracy and consistency influence decision quality jointly.

Cost and Value Management for Data Quality

The cost and value of data quality have been discussed in numerous articles; however, suitable and rigor cost measures and approaches to estimate the value are rare and indeed difficult to develop. At the same time, as a critical concern to the success of organizations, the cost and value of data quality become important. Numerous business initiatives have been delayed or even cancelled, citing poor-quality data as the main concern. Previous research and practice have indicated that understanding the cost and value of data quality is a critical step to the success of information systems. This chapter provides an overview of cost and value issues related to data quality. This includes data quality cost and value identification, classification, taxonomy, and evaluation framework, as well as analysis model. Furthermore, this chapter provides a guideline for cost and value analysis related to data quality.