Data acquisition and cost-effective predictive modeling: targeting offers for electronic commerce

摘要: Electronic commerce is revolutionizing the way we think about data modeling, by making it possible to integrate processes of (costly) acquisition and model induction. The opportunity for improving modeling through costly presents itself a diverse set electronic tasks, from personalization customer lifetime value modeling; illustrate with running example choosing offers display web-site visitors, which captures important aspects in familiar setting. Considering costs explicitly can allow building predictive models at significantly lower costs, modeler may be able improve performance via new sources information that previously were too expensive consider. However, existing techniques integrating cannot deal rich environment presents. We discuss several settings, challenges involved integration various research areas supply parts an ultimate solution. also present demonstrate briefly unified framework within one acquisitions different types, any cost structure objective.