Linguistic object-oriented web-usage mining Original Research Article

Web mining has become a very important research topic in the field of data mining due to the vast amount of world wide web services in recent years. The fuzzy and the object concepts have also been very popular and used in a variety of applications, especially for complex data description. This paper thus proposes a new fuzzy object-oriented web mining algorithm to derive fuzzy knowledge from object data log on web servers. Each web page itself is thought of as a class, and each web page browsed by a client is thought of as an instance. Instances with the same class (web page) may have different quantitative attribute values since they may appear in different clients. The proposed fuzzy mining algorithm can be divided into two main phases. The first phase is called the fuzzy intra-page mining phase, in which the linguistic large itemsets associated with the same classes (pages) but with different attributes are derived. Each linguistic large itemset found in this phase is then thought of as a composite item used in phase 2. The second phase is called the fuzzy inter-page mining phase, in which the large sequences are derived and used to represent the relationship among different web pages. Both the intra-page linguistic association rules and inter-page linguistic browsing patterns can thus be easily derived by the proposed algorithm at the same time. An example is given to illustrate the proposed algorithm. Experimental results also show the effects of the parameters used in the algorithm.