|
Activity Number:
|
339
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Risk Analysis
|
| Abstract - #308865 |
|
Title:
|
Using Data Mining Tools and GIS (Geographic Information Systems) in Hot-Spot Detection
|
|
Author(s):
|
Wei-hong Wang*+ and Ben Liu
|
|
Companies:
|
The College of New Jersey and William Paterson University
|
|
Address:
|
Department of Mathematics and Statistics, Ewing, NJ, 08628-4700,
|
|
Keywords:
|
Data mining ; Predictive modeling ; Neural Network ; GIS ; Decision Tree ; Logistic Regression
|
|
Abstract:
|
This study involves a specific data set which has 1.4 million cases, 14 predictors and one target variable with binary outcomes. The response surfaces and the classification accuracies of logistic regression and neural network models are relatively similar, yet the financial implications can be very profound when the models select the top 10% of cases and when the cost and profit are incorporated in the calculation. In addition, the Decision Tree and Geographic Information Systems (GIS) are used for case selection to enhance the profit to 96% of its full potential.
|
