Fully Nonparametric Models for Data Analysis: Some Case Studies
Paper
Fully Nonparametric Models for Data Analysis: Some Case Studies (.pdf document)
Program Files
*** Please read about the data sets before using.
For testing patterned alternatives
- data.txt: ASCII text file containing the teenage deviant behavior data
- analpa.q: Data analysis for data.txt
- alternatives tested: higher grades ->less marijuana use,
- higher parental education -> less marijuana use,
- females -> less marijuana use
- calibrate.q: Program used to calibrate bandwith for data.txt
Attached for omnibus test (a general alternative)
- analf.q: Data analysis for data.txt using finite-sample correction
- chisq.q: Data analysis for data.txt without finite-sample correction
- calibrate1.q: Bandwidth calibration for data.txt
All these programs only apply to the teenage deviant behavior data since the covariate is assumed to be integers between 4 and 21. This data has 8777 observations; by taking into account the covariate structure, the program saves a lot of time and memory. A general program is the following:
- tbbm.txt: Penn State Young Women's Health Study data
- tbbm.dat: Information on the dataset
- try.q: Contains the codes for calibrating the optimal bandwidth for tbbm.txt.
- try1.q: Contains the codes for calibrating the optimal bandwidth for tbbm.txt. Small-sample correction has been applied.
- ywhs.q: Data analysis for tbbm.txt without small sample corrections
- ywhs.sm.q: Data analysis for tbbm.txt with small sample corrections
Last modified: 06/04/08 | Contact Webmaster
