Hi Michael,
Thank you very much for replying immediately. I am still really impressed and excited about your informative and quick reply. This might perhaps not be a big deal for you but here in Germany I am not used to getting help and support from other researchers so instantly. I have had the experience that that many researchers here are not very approachable and too stressed to support and take care of other people’s questions and research problems. It’s a pity…So thanks again for your wonderful help. I cannot emphasize often enough how much I appreciate your help.
I have spent almost the entire last night on fully exploring NuMap since I had almost given up on using ANN for my purpose (I am simply interested in comparing the performance of a typical ANN with a multiple linear regression procedure from a very practical perspective – I won’t be able to go into ANNs theory too much). I have looked at number of freeware programs and some commercially available Demo versions (NeuroSolutions, Alyuda Forecaster XL, Easy NN, Tibera…) but all did not really satisfy me. They seemed either to be too complicated and confusing or too straightforward (not revealing what network topology, functions etc. are being employed) or documentation was missing.
NuMap seems to me the best structured program, is very user-friendly and extremely quick. I really enjoy using it and particularly like how the program takes you through the steps of preparing the network (Feature selection and Sizing) to Training and to finally Pruning and Processing the network. The only thing which I think could be improved is adding some more parameters for analyzing the performance of the network instead of only the MSE (e.g. coefficient of determination) and options for displaying some more graphs (e.g. predicted versus actual values). However, these things can also be obtained from using other available programs after exporting the data. So basically I came only across one minor thing with which I was struggling a little bit: when I exported data after “PROCESSING” the network the desired and actual results for one pattern are not displayed in two but only one column. But I used SPSS to restructure the! data. So actually not big deal either. To sum it up I am very happy to have found and got access to NuMap and now hope to be able to put my idea of using a MLP for predicting ****  (instead of using multiple regression analysis) into practice. This paper is supposed to be published in the Journal of applied Physiology. I will then be more than happy to acknowledge NuMap and your support certainly.
Thank you very much again in advance for helping me so much.
Kind regards
Alex
Alex***** @yahoo.com
Dear Michael Manry,
I have found your well-crafted NEUCL & NUMAP packages to be of great assistance in my philological research. Can you please quote the purchase price for the full package(s)? Thanks.
Regards,
Dan
Daniel M. Cro**
CRO**@ix.netcom.com
Dear Michael T. Manry
My name is Emad Elsa**. I am a postgraduate student. I use your superior package on my work for biomedical signals classification.
If you please, to send me more information about the Nearest Neighbour Classifier, since the available help in the package
is not enough. Thanks for all.Sincerely,
E. Elsa**
Dr. Manry,
…We are still making good use of the neural network package, thanks again, it’s a very easy and powerful tool.
Richard Od**
od**@compu**.com