Neural Network Modeling using SAS® Enterprise Miner Data Sets

Data Used to Compare Neural Network Estimates with Traditional Statistical Estimates

The following is the various SAS data sets that are used in Neural Network Modeling using SAS Enterprise Miner to compare neural network estimates with traditional statistical modeling estimates.

·

htwt.xls: This data was used to compare neural network estimates with regression modeling estimates. The school children data originates from the SAS/STAT User’s Guide Version 8 Example 55.2 Predicting Weight by Height and Age. 

·

penalty.xls: This data was used to compare neural network estimates with logistic regression regression estimates. The penalty data set is taken from Paul D. Allison book called Logistic Regression Using the SAS System: Theory and Application.

· delay.xls: This data was used to compare neural network estimates with nonlinear regression regression estimates. The radiation data set is taken from the SAS manual called SAS System for Regression by Rudolf Freund and Ramon Littell. The data represents radiation counts that were recorded nineteen separate times during a time span of four hundred hours.
· housecon.xls: This data was used to compare neural network estimates with time series  autoregressive forecasting estimates, that is, regression modeling with autoregressive errors. The construction data set is taken from the book called Forecasting Examples for Business and Economics Using the SAS System or SAS/ETS Software Applications Guide 1 from the SAS Institute. The time series data set consists of monthly construction contracts from January 1983 to October 1989 taken from the Survey of Current Business and Business Statistics.
· Leadprd.xls: This data was used to compare neural network estimates with one-step-ahead time series predictions. The lead production consists of monthly totals from a five-year period of U.S. lead production measured in tons, from January 1986 to September 1992. The lead production data set originates from the book called Forecasting Examples for Business and Economics Using the SAS System. 
· iris.xls: This data was used to compare the classification performance between the multivariate discriminant model with the neural network estimates. The classic iris data originates from Fisher 1936 that is listed in the SAS/STAT User’s Guide Version 8 Example 25.1: DISCRIM Procedure chapter.
Back to Previous Page

 

 

© copyright www.sasenterpriseminer.com - SAS neural network modeling.