An Association Rule-based Recommender Engine for Mobile Bird Searching Information Systems

Y.-P. Huang, W.-P. Chuang (Taiwan), and F.E. Sandnes (Norway)


Bird search system, Association rules, Recommendationmodel, Information retrieval.


In this paper, an intelligent bird search (IBS) system for mobile devices such as PDAs, Tablet PCs and smart phones is proposed. A user-friendly interface is designed that provides bird watchers with an easy and efficient way to identify an observed bird either graphically by making selections from a specific palette of feature icons or textually by keyword query. Once the target bird is found, the system provides detailed information about the species including multimedia information such as full color bird images, textual descriptions and digital audio recordings of bird sounds. Furthermore, users can add their own content as bird notes. This system has officially been demonstrated twice to the bird watchers in the national wild bird festivals held in 2002 and 2004 in Taiwan. The first demonstration was targeted at identifying the ease with which users could use the system by looking up entries and learn the information presented. This was achieved by logging the user behaviour in log files, including what features were used for the queries and what the target birds were. A minimum spanning tree and an Apriori-like algorithm were used to analyze users’ intentions and create a recommendation model for the IBS system. The recommendation model is designed from two viewpoints with the purpose to enhance the usability of the proposed IBS system. One is from the positive result and the other is from the null result. It also shows what species of wild birds are possible to flock together in the same area. During the second demonstration, the recommendation model was incorporated into the system and a post-test questionnaire was used to perform a quantitative user satisfaction survey.

Important Links:

Go Back