期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2012
卷号：42
期号：1
页码：100-112
出版社：Journal of Theoretical and Applied
摘要：Nowadays, most of e-service portals focus on individual users rather than considering shared characteristics that could be useful when searching a portal. Service providers searching activities involve different types of common user behaviors and preferences that should be considered. Thus, group profiling can be used to group customers with similar interests together and recommend service providers based on their common characteristics. Accordingly, this paper proposes a new approach that recommends service providers, based on group profiling and ranking, to be used by e-service portals. Such an approach is unsupervised and is carried-out in two steps. First, group profiles are generated. Second, whenever a user tries to search for a service provider, a different result may be returned based on the data collected about that specific user. Thus, customers would most likely find the most suitable service provider more accurately and in less time. This approach focuses on group profiling and group users based on the Ant Colony Clustering (ACC) method, which is a new type of clustering that is based on an Artificial Intelligence algorithm called Ant Colony Optimization. In addition, this approach has been tested and three measures have been recorded including speed, aggregation precision, and result accuracy. Such test has shown that this approach is promising and produces high quality results.
