Creating groups for marketing purposes from website usage data

Abstract

Customer grouping and knowledge extraction for these groups are important to online businesses because it allows purposeful application of marketing techniques. Individuals can be personally served with the groups, depending on the identified interests and preferences. In this article, we suggest a way to identify and create user groups by processing website usage data. We use the logs stored in the server log data for the visit to a selected website and then retrieve and process the text content of the visited web pages. The approach is based on the technology for natural language processing and uses the methods for clustering of text documents. The experimental testing of this method is done with the software product RapidMiner and data from visits to a Bulgarian e-shop.