Using multiple modeling techniques on the same data set

Conclusions and Implications

The purpose of this White Paper is not to push one technique over another. We did not even get into a discussion of other modeling techniques, such as neural networks, the complexity of which is beyond the intentionally non-technical level of this discussion.

And we did not cover issues related to model validation testing, which is beyond our intended scope. (For simplicity, we have assumed in the above discussion that all three models will perform comparably in validation tests.)

We just wanted to expose the reader to some issues related to various modeling techniques, and to show that sometimes it is useful to try more than one technique in the same data mining project. (If we had used a different data set, the comparison of techniques might have told a different story!)

An eclectic approach, along with careful examination and application of the results, can lead to more successful and profitable modeling outcomes.

				Conclusions and Implications
Linear Regression Model	Logistic Regression Model	CHAID Segmentation Model	Comparison of CHAID vs. Regression Models