Community Profile

Greg Heath


7536 total contributions since 2004

Backgound in Electromagnetic Theory, Plasma Physics and Radar Target Identification using Neural Networks.
PhD Student, Research Assistant and Lecturer at Stanford;
AB,ScB, ScM Student, Research Assistant, Fellow and Professor at Brown;
27 yrs researching Ballistic and Theatre Missile Defense using Neural Networks at MIT Lincoln Laboratory. Retired 2003.

PLEASE DO NOT SEND QUESTIONS AND DATA TO MY EMAIL. HOWEVER, CAN SEND LINKS TO POSTS.
Professional Interests: Neural Netwoks, Spectral Analysis

Contact

Contributions in
View by

Responded


Re: Large negative output values with tanh activation for neural networks
This is not the same as Hope this helps. Greg .

1 day ago | 108 views

Answered
how to get image after training in a neural network
Your output will be in matrix form which can be converted to an image. The type of image will determine the type of conversio...

1 day ago | 0

| accepted

Answered
i want use multi layer Perceptron not using nntool using this code
After reading the EXCEL files you have to convert them to MATLAB matrices with sizes [ I N ] = size(input) [ O N ] = siz...

1 day ago | 0

Answered
How can i get a classifier at the neural network toolbox in matlab r2015b
Regression net: help fitnet doc fitnet Classifier net: help patternnet doc patternnet ALSO: You can search the...

1 day ago | 0

Answered
future value prediction using narx
Scanty info. [ Xc Xci Aci Tc ] = preparets(netc,X,{},T); [ Yc Xcf Acf ] = net(Xc, Xci, Aci); where (Xci,Aci) and (Xcf...

4 days ago | 0

| accepted

Answered
Confusion matrix outputs NaN values after classification using Neural Network
Isn't it obvious? 1. Rows 9 and 10 contain nothing but zeros. 2. Ratios and percentages obtained by dividing by zero w...

4 days ago | 0

Answered
How to disable popup of training windows for neural network in matlab
close all, clear all, clc [x,t] = simplefit_dataset; net = fitnet; net.trainParam.showWindow = 0; % <== This does it [n...

9 days ago | 0

Answered
How to develop a neural network transfer function?
Use FITNET with vectors obtained from vectorizing layers. Hope this helps *Thank you for formally accepting my answer* ...

9 days ago | 0

| accepted

Answered
How to synthetic generate features to balance dataset?
If you have multiple classes with different numbers of examples, the means and covariances of the classes can be used to generat...

9 days ago | 0

Answered
how to get the best the number of hidden neurons and layers for a NARXNET?
There are zillions of my posts in both the NEWSGROUP and ANSWERS that use the criterion Ntrneq >= Nw %( number of...

9 days ago | 0

Answered
Train recurrent neural network with variable length input cell vectors.
Neural networks require "I"nput column vectors with a fixed length (I). Therefore you will have to complete the columns with zer...

9 days ago | 0

Answered
How can I manually perform an elmannet neural network calculation?
My guess is z(t) = B1 + IW * [ x(t); z(t-1); z(t-2)]; y(t) = B2 + LW * z(t); Hope this helps *Thank you for formally...

10 days ago | 0

| accepted

Answered
Where can I find the coefficients for my neural net input elements?
Typically, the output y is given by y = B2 + LW*tanh( B1 + IW*x) where B1 input "B"ias net.b{1} IW ...

11 days ago | 0

| accepted

Answered
How can train a neural network, using Genetic Algorithm, I need a example in code. thank
Search NEWSGROUP and ANSWERS with greg genetic Hope this helps. *Thank you for formally accepting my answer* Greg

11 days ago | 0

Answered
Is possible solve the problem used ANN toolbox? (related future input)
"I"nputs are I-dimensional column vectors and "O"utput targets are O-dimensional column vectors. Then [ I N ] = size(inputmat...

11 days ago | 0

Answered
neural network input error
You cannot change the dimension of the inputs from 50 to 100; Hope this helps *Thank you for formally accepting my answer*...

11 days ago | 0

Answered
Neural Net Fitting - how to set a result goal?
I'm sure this has been discussed in both NEWSGROUP & ANSWERS at least a zillion times Try searching with greg msegoal ...

11 days ago | 0

Answered
How to train ANN if my input data is in the form of signals (EMTP-ATP)?
There are two separate regions with stationary summary statistics. This can be modeled with two nets. Why does the secon...

11 days ago | 0

Responded


Re: Question on Data Division Techniques in Neural Networks
Sounds good. However, this can be done with either one or two initializations. Would have to dig into the code to be sure. Hop...

11 days ago | 296 views

Answered
Importance of various inputs of neural network
I don't think the analytic expression y = B2 + LW * tanh( B1 + IW * x ) will help very much Ranking a large number of ...

14 days ago | 0

Responded


Re: Question on Data Division Techniques in Neural Networks
I don't know the exact reason. However, I do know that when you use a rng to generate random values, you only initialize the rng...

14 days ago | 296 views

Responded


Re: Bug in Neural Net calcs for log sigmoid?
close all, clear all, clc, format compact rng(0) %FOR REPRODUCIBILITY p = [ 0 1 0 1 ; 0 0 1 1 ]; t =...

16 days ago | 179 views

Answered
How to get the transfer function equation for the implemented neural network
You have to obtain xsettings and tsettings from x and t. Then apply them to x2 and t2. Hope this helps. *Thank you for ...

1 month ago | 0

| accepted

Answered
How do I get the correct output from a fitnet artificial neural network?
1. Very often you have to use several choices of the random initial weights in order to get a good answer. 2. You forgot to t...

1 month ago | 0

| accepted

Answered
In neural network toolbox for the function: trainautencoder, how can I use purelin for encodertranferfunction?
Using purelin in a hidden layer is a waste of time. You can get the same result by removing the layer and adjusting the outpu...

1 month ago | 0

| accepted

Answered
In Neural Network Toolbox, how can I can change the values of training Parameters ?
1. Another way to do it is net = patternnet(hiddenlayersize); . . . net.trainFcn = 'traingdx'; [ net tr y e ] = train(n...

1 month ago | 0

Answered
Coding ANN learning parameters (e.g number of hidden neurons, learning methods, learning rate, momentum, number of epoch) in Genetic Algorithm chromosome?
Search both NEWSGROUP and ANSWERS using greg genetic Hope this helps *Thank you for formally accepting my answer* ...

1 month ago | 0

Answered
How to change learning rate and momentum rate values in GUI application to train ANN
Use the default values and worry about 1. Number of hidden nodes 2. Initial random weights For examples search in both ...

1 month ago | 0

Answered
in backpropagation learning algorithm always i used to get the same accuracy rate whatever learning rate and momentum constant is used.So, what is the solution for this
I have the same experience. What I do is use the best of multiple designs that differ by number of hidden nodes and initial r...

1 month ago | 0

Load more