Proposing a classification algorithm for user identification according to user web log analysis

Request classification is one of important strategies of Web. User behavior analysis can make Web service more intelligent and secure. Recently, tree structures have become a popular way for storing and manipulating huge amount of data. The classification of these data can facilitate storage, retrieval, indexing, query answering and different processing operations. In this paper, we propose User- Classifier algorithm for rule based classification of tree structured data that to classify web log user. This algorithm is based on extracting special tree pattern from training dataset. Our experiments show that User- Classifier reduces running time. In the case of complete classification, User- Classifier shows the best classification quality.