Rule Optimization of Web-Logs Data Using Evolutionary Technique

2. Related Works

Web usage mining consists of three phases: preprocessing, pattern discovery, and pattern analysis. B. Santhosh and Rukmani(2010) worked on ‘Implementation of Web Usage Mining by using Apriori and FP-growth Algorithms’. Authors used Apriori algorithm to generate association rules that identifies the usage pattern of the client for a particular website. The output of the system is in term of memory usage and speed of producing association rules. Iyakutty and Sujatha (2010) proposed a new framework for Web usage data clustering for user’s session. Web clustering involves grouping of the similar object and dissimilar object in different group. The initial clusters are selected based on statistical model to allow the iterative algorithm to converge to a better local minima and improving cluster quality using genetic algorithm based refinement. The method is scalable and can be coupled with a scalable clustering algorithm to address the large-scale clustering problems in Web mining. Biwei Li and Cunlai Chai (2010) presented a GA-based method to derive the fuzzy sets from a set of given transactions. Genetic algorithms provide efficient search algorithms to select a model, from mixed media data, based on preference criterion and objective function. It combines the strengths of rough set theory and genetic algorithm. Arslan et al. (2006) proposed method to find sequential accesses from weblog files using genetic algorithm. Weblog transaction, whether completed or not, is recorded and stored unstructured. Analyzing these log files is one of the important research areas of Web mining. Gyenesei (2000) presented methods for mining fuzzy quantitative association rules; namely without normalization and with normalization. The results showed that the numbers of large itemset and interesting rules found by fuzzy method are larger than the discrete method (Agrawal and Srikant, 1999). Hadzic and Hecker, (2011) presented an approach where a tree structured data is converted into flat represenattion for preserving the structural and attribute value information, thus enabling a wider range of data mining and analysis techniques. Luan et al. (2012) introduced an association rule algorithm for Web log mining that reduces the search range and avoids the problem of combinatorial explosion. Nithya and Sumathi (2012) focused on data cleaning by removing the noise. Weber et al. (2012) suggested an approach where data blogs can be used to visualize political issues covering various sub issues. In Mele (2013), Author focused on improving the search engine performance by using static caching and recommending the interesting Web pages, articles and blogs.