Recommendation Approaches Based On Web Usage Mining

With the rapid growth in Internet and WWW, the user'browsing information is becoming enormous and pervasive, which represents the user access details with the user dimension, time dimension, space dimension, and access object dimension. Mining the user access information, we can obtain the knowledge about user access manners, which can be used for the service providers and users.For the service providers, they require the good automatic assistant design tools that can dynamically adjust web topology, improve service, and carry out personalization electronic commerce for users according to the user access interest knowledge. For users, they need the personalization web pages and special service and get the valuable recommendation from others who have similar interest. Therefore, how to automatically and effectively get the knowledge from the vast user access information, i.e. web usage mining, became a new and important research field in the world.This dissertation addresses the researches of the web adaptive research field and the personalization research field in web usage mining. The contribution of the dissertation is as follow:1. The web adaptive research field:1) Clustering on web broadcast: To resolve the information broadcast problems in broadband network, this dissertation presents a new web mining approach– WebClustering, which can make a new and valuable web broadcast set, and some layered index web pages generated to help users access the set.2) Discovering large sequences and mutual information rules:To mine the large sequences in user access sequences, this dissertation presents an approach that discovers the large sequence. It defines a kind of the new user access transaction grammar to get the user sequence access transactions from the user access transactions. To mine the mutuality themes, the mutual information discovering approach is used and a new clustering approach is given. The knowledge can help the web site designers deeply understand users'access in order to adjust the web site structures.