Accessibility navigation

Solving the stability-accuracy-diversity dilemma of recommender systems

Liu, K., Hou, L., Liu, J. and Zhang, R. (2017) Solving the stability-accuracy-diversity dilemma of recommender systems. Physica A: Statistical Mechanics and its Applications, 468. pp. 415-424. ISSN 0378-4371

Text - Accepted Version
· Please see our End User Agreement before downloading.


It is advisable to refer to the publisher's version if you intend to cite from this work. See Guidance on citing.

To link to this item DOI: 10.1016/j.physa.2016.10.083


Recommender systems are of great significance in predicting the potential interesting items based on the target user’s historical selections. However, the recommendation list for a specific user has been found changing vastly when the system changes, due to the unstable quantification of item similarities, which is defined as the recommendation stability problem. To improve the similarity stability and recommendation stability is crucial for the user experience enhancement and the better understanding of user interests. While the stability as well as accuracy of recommendation could be guaranteed by recommending only popular items, studies have been addressing the necessity of diversity which requires the system to recommend unpopular items. By ranking the similarities in terms of stability and considering only the most stable ones, we present a top-n-stability method based on the Heat Conduction algorithm (denoted as TNS-HC henceforth) for solving the stability-accuracy-diversity dilemma. Experiments on four benchmark data sets indicate that the TNS-HC algorithm could significantly improve the recommendation stability and accuracy simultaneously and still retain the high-diversity nature of the Heat Conduction algorithm. Furthermore, we compare the performance of the TNS-HC algorithm with a number of benchmark recommendation algorithms. The result suggests that the TNS-HC algorithm is more efficient in solving the stability-accuracy-diversity triple dilemma of recommender systems.

Item Type:Article
Divisions:Henley Business School > Business Informatics, Systems and Accounting
ID Code:68450


Downloads per month over past year

University Staff: Request a correction | Centaur Editors: Update this record

Page navigation