search for: prediciton

hello, I apologize that I poste this question again but I did not receive any replies on this topic and would like to know if the mail has been read over or if there is no method of statistical prediction computation for poisson errors in glm(). I am looking for a way to compute the model prediciton error of a glm() with poisson error family. cv.glm() does not work. (it prompts values around 90.00) A " not there isn't such a method" would be fine as well. Thanks in advance, Martin

Hi! In a current project, I am fitting loess models to subsets of data in order to use the loess predicitons for normalization (similar to what is done in many microarray analyses). While working on this I ran into a problem when I tried to predict from the loess models and the data contained NAs or NaNs. I tracked down the problem to the fact that predict.loess will not return a value at all when fed wi...

Dear all, I have got two vectors coding for a stimulus presented in the current trial (mydat$Stimulus) and a prediction in the same trial (mydat$Prediciton), respectively. By applying an if-conditional I want to create a new vector that indicates if there is a match between both vectors in the same trial. That is, if the prediction equals the stimulus. When I pick out some trials randomly, I get some trials with no match (mydat$Stimulus[1] != mydat$P...

...s to Cache the file and use the old incorrect data. If you modify the file with a PC application then everything is fine and the "Cache" is updated. I was unaware that Samba was caching these files if indeed that is what is going on. I have not got the getwd cache set at all or the read prediciton set either. Since they are not set in the smb.conf file I assume that they are set to no. Even if you reboot the PC the file is not refreshed and you get the old file back even though it has been modified by another UNIX client. How should I have my locking options set I tend to think that my prob...

Hi all, I ran a discriminant analysis with some data and want to get a general idea of prediction error rate. Some have suggested using X-fold cross validation procedure. Anyone know if there is a function for this in R? Thanks, Wade [[alternative HTML version deleted]]