A recommendation model for social resource sharing systems based on tripartite graph clustering

The use of folksonomies to recommend web pages and tags assigned to these pages, is an important research direction in web recommendation. In this study, we implement a model that fits tripartite structure of folksonomies and extracts valuable information for generating recommendations. Then we developed two types of recommendation systems that take advantage of this information; web page recommendation and tag recommendation. We compared our recommendation results with the results using bipartite clustering of web pages and tags. The experiments are conducted on the data set obtained from Del.ici.ous web site. The results show that this model generates better accuracy results for web page recommendation while extracting more useful information simultaneously which could be an extra to generate different types of recommendations.