search for: ssden

...covariates, going beyond the proportional hazard models. Utilities are provided for the calculation of a certain Kullback-Leibler projection of cross-validated fits to "reduced model" spaces, for the "testing" of model terms. Projection code is provide for ssanova1, gssanova1, ssden, and sshzd fits. Further details are to be found in the documentations and the examples therein. As always, feature suggestions and bug reports will be sincerely appreciated. Chong Gu

...covariates, going beyond the proportional hazard models. Utilities are provided for the calculation of a certain Kullback-Leibler projection of cross-validated fits to "reduced model" spaces, for the "testing" of model terms. Projection code is provide for ssanova1, gssanova1, ssden, and sshzd fits. Further details are to be found in the documentations and the examples therein. As always, feature suggestions and bug reports will be sincerely appreciated. Chong Gu

Hi, I need to estimate the density at the mean of a sample of a few thousands data points with a dimesion up to 5. The data is uni-modal and regularly shaped. I couldn't find any kernel density package for R which supports more than 3 dimensions. Have I overlooked a package or does somebody have code for this purpose? Any other advice? Regards, Stephan

Hello: I have been using sm.binomial() in the Bowman and Azzalini's sm package to smooth and plot estimated probabilities as a function of a covariate. I am concerned about my choice of bandwidth, and I was hoping there was another method available in some other package, perhaps with an automatic choice of smoothing parameter. Does anyone know of one? Thanks in advance. Tom Richards