Evaluating collaborative filtering recommender systems

Herlocker, Jonathan L.; Konstan, Joseph A.; Terveen, Loren; Riedl, John

doi:10.1145/963770.963772

Cited by 4,669 publications

(2,992 citation statements)

References 20 publications

Supporting

Mentioning

2,751

Contrasting

Unclassified

100

Order By: Relevance

“…Users have to move an arrow showing their degree of similarity with the personality characteristics of the characters being presented (these personality characteristics along with examples of typical behaviour patterns are presented under each character's image). In [13] we concluded that it is possible to replace the TKI personality test with the movie metaphor test because it provides an statistically The results that the TKI test provided for the five different personality modes in comparison with the values that the movie metaphor test gathered had a Mean Absolute Error (MAE) [52] of 0.12. Hence, we believe that it is worth sacrificing a little accuracy in the test results (as they are not for psychology testing purposes) in exchange of enhancing significantly the usability and interest for the application.…”

Section: Social Knowledge Elicitation Modulesmentioning

confidence: 99%

Development of a group recommender application in a Social Network

Quijano-Sánchez

Díaz-Agudo

Recio-García

2014

Knowledge-Based Systems

View full text Add to dashboard Cite

In today's society, recommendations are becoming increasingly important. With the advent of the Social Web and the growing popularity of Social Networks, where users explicitly provide personal information and interact with others and the system, it is becoming clear that the key for the success of recommendations is to develop new strategies which focus on social recommendations leveraged by these new sources of knowledge. In our work, we focus on group recommender systems. These systems traditionally suffer from a number of shortcomings that hamper their effectiveness. In this paper we continue our research, that focuses on improving the overall quality of group recommendations through the addition of social knowledge to existing recommendation strategies. To do so, we use the information stored in Social Networks to elicit social factors following two approaches: the cognitive modeling approach, that studies how people's way of thinking predisposes their actions; and the social approach, that studies how people's relationships predispose their actions. We show the value of using models of social cognition extracted from Social Networks in group recommender systems through the instantiation of our model into a real-life Facebook movie recommender application.

show abstract

Section: Social Knowledge Elicitation Modulesmentioning

confidence: 99%

Development of a group recommender application in a Social Network

Quijano-Sánchez

Díaz-Agudo

Recio-García

2014

Knowledge-Based Systems

View full text Add to dashboard Cite

show abstract

“…CBF approaches and CF algorithms have both been used fairly successfully to build recommendation systems in various domains [11][12][13]16]. However, as described above, they suffer from the cold-start problem in one form or another.…”

Section: Related Workmentioning

confidence: 99%

A Content-Boosted Collaborative Filtering Algorithm for Personalized Training in Interpretation of Radiological Imaging

2014

View full text Add to dashboard Cite

Devising a method that can select cases based on the performance levels of trainees and the characteristics of cases is essential for developing a personalized training program in radiology education. In this paper, we propose a novel hybrid prediction algorithm called content-boosted collaborative filtering (CBCF) to predict the difficulty level of each case for each trainee. The CBCF utilizes a content-based filtering (CBF) method to enhance existing trainee-case ratings data and then provides final predictions through a collaborative filtering (CF) algorithm. The CBCF algorithm incorporates the advantages of both CBF and CF, while not inheriting the disadvantages of either. The CBCF method is compared with the pure CBF and pure CF approaches using three datasets. The experimental data are then evaluated in terms of the MAE metric. Our experimental results show that the CBCF outperforms the pure CBF and CF methods by 13.33 and 12.17 %, respectively, in terms of prediction precision. This also suggests that the CBCF can be used in the development of personalized training systems in radiology education.

show abstract

“…Here, we first use Rank Score [15] to measure the ability of a recommendation algorithm to produce a good ordering of items that matches the user's preference. Since real users usually consider only the top part of the recommendation list, two more practical measures are also be employed, namely Precision and Recall [20].…”

Section: Data and Metricsmentioning

confidence: 99%

Preference of online users and personalized recommendations

Yuan

Zhao

Zeng

et al. 2013

Physica A: Statistical Mechanics and its Applications

View full text Add to dashboard Cite

In a recent work [Proc. Natl. Acad. Sci. 107 (2010) 4511], a personalized recommendation algorithm with high performance in both accuracy and diversity is proposed. This method is based on the hybridization of two single algorithms called probability spreading and heat conduction, which respectively are inclined to recommend popular and unpopular products. With a tunable parameter, an optimal balance between these two algorithms in system level is obtained. In this paper, we apply this hybrid method in individual level, namely each user has his/her own personalized hybrid parameter to adjust. Interestingly, we find that users are quite different in personalized hybrid parameters and the recommendation performance can be significantly improved if each user is assigned with his/her optimal personalized hybrid parameter. Furthermore, we find that users' personalized parameters are negatively correlated with users' degree but positively correlated with the average degree of the items collected by each user. With these understandings, we propose a strategy to assign users with suitable personalized parameters, which leads to a further improvement of the original hybrid method. Finally, our work highlights the importance of considering the heterogeneity of users in recommendation.

show abstract

Evaluating collaborative filtering recommender systems

Cited by 4,669 publications

References 20 publications

Development of a group recommender application in a Social Network

Development of a group recommender application in a Social Network

A Content-Boosted Collaborative Filtering Algorithm for Personalized Training in Interpretation of Radiological Imaging

Preference of online users and personalized recommendations

Contact Info

Product

Resources

About