Extraction of relevant information from the web has become the emerging challenge as the data in World Wide Web has been progressively increased. To solve this problem, a valid method called web usage mining is used that can mine weblogs for user models and recommendations. The general recommender systems of the web mainly function on the basis of association rule mining and clustering. Apart from this, web personalization is proposed in this paper which adopts sequential access pattern mining. The recurrent sequential web access patterns are recognized through this method. Those patterns are further stored in tree structure termed as pattern-tree. The subsequent processes such as matching and producing web links for recommendation are done with these stored patterns. Finally, a valuable performance analysis is carried out to validate the proposed model.
