查看文章

Nowadays, providing tools that eases the interaction of users with websites is a big challenge in e-commerce. Recommender systems are useful tools which adapts the environment of websites compatible with users needs. In this paper, applying a hybrid collaboration and content based technique a model for recommendation system is proposed. Presented model works in two offline and online phases. In offline step the behavior of users’ models with a combined FCM and ant based clustering algorithm and in online step suitable recommendations extracts for presenting to active user. The model is implemented and tested as a recommender system for personalizing website of “Information and Communication Technology Center” of Isfahan municipality in Iran. The results shown are promising and proved that applying more efficient clustering technique for modeling users behavior provide us with more interesting and useful patterns which consequently making the recommender system more functional and robust.