Siddhartha Saha <[EMAIL PROTECTED]> wrote in message a72i9l$huhbg$[EMAIL PROTECTED]">news:a72i9l$huhbg$[EMAIL PROTECTED]... > Hi, > I am basically a CS guy, and I am trying to compare a neural network based > approach of solving the problem below to some statistical approach. It would > be nice if someone could provide me some help on the latter part. > > I have some data at hand corresponding to signal strength values at 3 > different channels at some given 19-20 locations. For each location, I have, > say a few 100 samples at successive sampling intervals. Due to the nature of > the sampling it cant be said for sure that each sample is independent. But > the channels are from independent sources. This constitutes my training set. > Now my test set is another set of signal strength samples, taken at one of > these locations. I would like to perform some hypothesis test to determine > which location the data came from. What would be a appropriate test to do > this?
That sounds like a classification problem, not a hypothesis test problem. Why would you use testing rather than what is effectively estimation? Glen . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
