The Role of Domain Knowledge in KDD-based Strategic Marketing Applications

A key feature of knowledge discovery based systems is the ability to keep into account the knowledge that users own about the application fields (domain knowledge). Such kind of ability needs a strong multidisciplinary approach to ensure the novelty of discovered knowledge and its understandability in terms of a user meaningful model. Traditional strategic marketing decision support systems, in order to provide trustful results, are usually founded on sound economic theories that represent their background domain knowledge. However, KDD-based marketing applications, aiming at exploiting enterprise information asset, are designed understating the role of business background knowledge. It leads to build, despite their accuracy, models that aren’t able to explain market dynamics, in which strategic marketing users are mainly interested. The inability of KDD systems to accommodate domain knowledge limits their application fields. In order to effective combining KDD and economic modeling in building a strategic marketing support system that exploits enterprise data, we have designed a methodology that integrates contributions from both Microeconomic and Data Mining fields. In this paper we present our framework in terms of interdisciplinary associations and an application to the churn analysis in the Italian telecommunication market.