Compute Normal Vector

compute.nv(xmat, resp, nv.method, ...)

Arguments

xmat	Matrix of OC coordinates (i.e., predictors).
resp	Response Variable (i.e., ordered choices).
nv.method	Method of normal vector estimation. Following pre-defined methods are currently available: `"oprobit"`: Ordered probit regression (calls `polr` function from `MASS` package) `"ologit"`: Ordered logistic regression (calls `polr` function from `MASS` package) `"svm.reg"`: Support Vector Regression (calls `nv.svm` function) `"svm.class"`: Support Vector Classification with dichotomized response (calls `nv.svm` function) `"krls"`: Kernel Regularized Least Squares (calls `nv.krls` function). Alternatively, one can set custom made function that (at least) feeds in `xmat` and `resp` arguments and returns the vector of coefficients (the length must be equal to the number of columns of `xmat`).
...	Additional arguments passed to the function assigned by `nv.method`.

The vector of normal vector estimates.