Extraction of Target User Group from Web Usage Data Using Evolutionary Biclustering Approach

Extraction of Target User Group from Web Usage Data Using Evolutionary Biclustering Approach

R. Rathipriya (Periyar University, India), K. Thangavel (Periyar University, India) and J. Bagyamani (Government Arts College, India)
DOI: 10.4018/978-1-4666-2145-9.ch015
OnDemand PDF Download:
List Price: $37.50


Data mining extracts hidden information from a database that the user did not know existed. Biclustering is one of the data mining technique which helps marketing user to target marketing campaigns more accurately and to align campaigns more closely with the needs, wants, and attitudes of customers and prospects. The biclustering results can be tuned to find users’ browsing patterns relevant to current business problems. This paper presents a new application of biclustering to web usage data using a combination of heuristics and meta-heuristics algorithms. Two-way K-means clustering is used to generate the seeds from preprocessed web usage data, Greedy Heuristic is used iteratively to refine a set of seeds, which is fast but often yield local optimal solutions. In this paper, Genetic Algorithm is used as a global optimizer that can be coupled with greedy method to identify the global optimal target user groups based on their coherent browsing pattern. The performance of the proposed work is evaluated by conducting experiment on the msnbc, a clickstream dataset from UCI repository. Results show that the proposed work performs well in extracting optimal target users groups from the web usage data which can be used for focalized marketing campaigns.
Chapter Preview

Tsai et al. (2004) developed a market segmentation methodology based on product specific variables such as purchased items and the associative monetary expenses from the transactional history of customers to address the unreliable results of segmentation based on general variables like customer demographics. Shina et al. (2004) used three clustering methods such as K-Means, Self-Organizing Map, and fuzzy K-Means for segmentation to find properly graded stock market brokerage commission rates based on transactional data. Punj and Stewart (1983), suggest two approaches for clustering. The first was hierarchical clustering, to determine the number of clusters, and the second was nonhierarchical clustering, for fine-tuning the results. In which, hierarchical cluster analysis was extremely time consuming, it is rarely used in practice. Jonker et al. (2004) used the Genetic Algorithm (GA) to determine the optimized marketing strategy to integrate customer segmentation and customer targeting. In Abraham (2003), intelligent-miner” (i-Miner) is introduced and optimized concurrent architecture of a fuzzy clustering algorithm is used to discover web data clusters and a fuzzy inference system to analyze the trends of the web site visitors.

Complete Chapter List

Search this Book: