rcmreg - computers the regression coefficient estimates based on the lift rank covariance matrix
rcmreg [-f filename -n obs -k vars -y resps
[-b nrboots] ]
rcmreg computes the regression coefficient estimates using the lift rank covariance matrix, where the multivariate rank concepts are based on the Oja criterion function. User may invoke a resampling strategy to estimate the standard errors of the estimates. When no command line parameters are given, the program asks the parameters from the user.
-f filename
specify the file filename as the data file
-n obs
specify obs as the number of observations in the data.
-k vars
specify vars as the number of variables.
-y resps
specify resps as the number of response variables. The response variables are expected to be the last columns in the given data file filename.
-b nrboots
use this option if you wish to invoke bootstrap procedure to estimate the standard errors of the regression estimates. Parameter nrboots specifies the number of bootstrap samples.
example1: rcmreg -f mydata.dat -n 25 -k 4 -y 2 -b 200
The file containing the data is called mydata.dat. There is 25 observations and 4 variables in the data. The response variable is of dimension 2 so there are two explanatory variables. (The response variables are supposed to be the last columns in the data file). Bootstrap procedure with 200 bootstrap samples is invoked.
See paper:
E. Ollila, H. Oja and V. Koivunen, ``Estimates of regression coefficients based on lift rank covariance matrix,'' Journal of the American Statistical Association, vol. 98, no. 461, pp. 90-98, 2003.
Esa Ollila, esollila AT wooster DOT hut DOT fi
Please send any comments, bug reports, suggestions, etc to the above email address. All comments are highly appreciated.