ABSTRACT
As with any information technology development project, executive sponsorship can be a key element in attaining the necessary resources for a successful project and the commensurate recognition, which in turn enhances opportunities to proceed with subsequent data-mining activities. Based on conversations with financial institution executives, several areas may be identified for analysis. It is recommended that the highest-priority subject area be selected as the initial focus of the data-mining project, in order to provide meaningful results from the activity. The target area could be customer acquisition profiling, customer retention, fraud detection, liability prediction, or virtually any area where sufficient historical data exist to support the analysis. The techniques that can be implemented as software routines for use in data mining can be grouped according to function: link analysis, classification, clustering, statistics, prediction. These techniques are described, along with specific techniques that are useful in performing each function.
