Data mining case studies in customer profiling

The purpose of this document is to report the results of applying some data mining methods that were reviewed in the earlier customer profiling report, to real world problems. The cases in question concern customer data from a bank and a book club. Two separate approaches are taken in the problem solving. First, an explorative analysis is applied to both cases. The goal of the analysis, as can be expected, is to give insight of the data in order to reveal or extract the relevant structure of and patterns in the data. The specific tasks include clustering of the banking data and sequence mining of the book club transactions. Subsequently, a predictive analysis of the cases is performed. Unlike the explorative analysis, the prediction involves a target variable to be considered. The general aim of the analysis is to generate a model using examples of the past in order to predict the outcome in the future. The concrete tasks performed are to predict the profitability of the book club members with logistic