Regression Models for Predicting Building Material Distribution in Four Northeastern Cities.
Abstract
The Corps of Engineers conducted a field sampling program for inventorying building materials in the northeastern United States, and the data from the field program were compiled into a data base for statistical analysis. Correlation coefficients were derived between the independent variables and the surface area of the five building material types. The correlation coefficients were used in an optimal stepwise regression model developed for each material class for each city. A number of factors appear to be significantly associated with the distribution of building material exposure. However, the variables do not correlate at levels required for constructing adequate predictive models that would be applicable to other sampling locations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1985
- Accession Number
- ADA166335
Entities
People
- Carolyn J. Merry
- Perry J. Lapotin
Organizations
- Cold Regions Research and Engineering Laboratory