36
■
Chapter 2 Principal Model Types
can generate rules that produce a model prediction. After all, a neural
network that learns to classify credit applicants into Good, Review, and
Reject is making a prediction about the probable future fate of a loan.
2.3 Review
From an understanding of distributed intelligence and data mining in
Chapter 1, this brief chapter provides a review of the various forms a
model can take. Model building is more than applying statistics, creating
rules, or generating a decision tree. Model building involves understand-
ing the nature of the problem and finding the system representation that
best reflects the expectations of the consumer, the capabilities of the sys-
tems engineer, and the availability of supporting data. From ...