Community Profile

photo

Greg Heath


3,103 2011 年以降の合計貢献数

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

連絡する

Greg Heath's バッジ

  • 36 Month Streak
  • Thankful Level 4
  • Ace
  • Revival Level 4
  • Knowledgeable Level 4
  • First Answer

詳細を表示...

貢献した分野
表示方法

回答済み
Hyperparameter tuning of neural network
One hidden layer is always sufficient. However, sometimes 1. Knowledge of the physical or mathematical process may lead to a ...

1日 前 | 1

回答済み
Different results in training a CNN with Matlab 2018a and Matlab 2019a
You are making the task difficult by going backwards. Start with a single hidden node and add nodes one at a time. Hope this...

3日 前 | 0

質問


NEURAL NETWORK DATA SET EXAMPLES
For demonstration of old AND new concepts and ideas, PLEASE use the sample NN data sets provided by MATLAB help nndatas...

3日 前 | 0 件の回答 | 0

0

回答

回答済み
cross validation in neural network using K-fold
%i am using neural network for classification but i need to use instead of holdout option , K-fold. ==> FALSE!. You mean y...

5日 前 | 0

回答済み
Can the number of Predictors be different for Train and Test data?
Of course not. The ultimate purpose of training is to create a model that works well on non-training data. Thank you for form...

5日 前 | 0

回答済み
How to check the robustness of the Neural network model?
If you are going to test with white noise, include white noise in your design (i.e., training + validation) Then, given a fixed...

9日 前 | 0

| 採用済み

回答済み
NARX with Complex Values Input
Decades ago I learned (the hard way) to forget about trying to use complex computations for NNs. However, if you insist, let us...

19日 前 | 1

回答済み
Why sets Matlab automatically the activation functions for a neural network like this?
That is a standard configuation for a neural net. It's operation is explained in every elementary text. Thank you for formally...

24日 前 | 0

回答済み
I get a "Performance function replaced with squared error performance" warning when trying to set 'crossentropy' as the performance function.
If you insist on using CROSSENTROPY, try PATTERNNET. Hope this helps. Thank you for formally accepting my answer Greg

約1ヶ月 前 | 0

回答済み
How to plot Network performance?
You have lost training information, So the only thing left is output vs input. Hope this helps. Thank you for formally ac...

約1ヶ月 前 | 0

回答済み
Elman Neural Network (ENN)
size(P_TRAIN) = [ 1296 1728] size(T_TRAIN) = [ 432 1728] Hope this helps. *Thank you for formally accepting my answer* ...

約1ヶ月 前 | 0

回答済み
NARX re-training in closed loop
Using 100 feedback delays makes no sense. Only use feedback delays that are within the correlation length of the function. See...

約2ヶ月 前 | 1

回答済み
Can CNN train separately instead of learning everything at one time?
I try to have the order of inputs as random and uncorrelated as possible. Otherwise the probability of extensive learning/unlear...

約2ヶ月 前 | 0

回答済み
In evaluating a neural net, should NMSE be based only on test subset of data?
For serious work I calulate FOUR values of NMSE: 1.70% Training 2.15% Validation 3.15% Test 4.100% All for 10 (typically...

2ヶ月 前 | 0

| 採用済み

回答済み
Is the "patternnet" a fully connected neural network
Yes. The only difference between my classifiers and regressors is the sigmoid output layer instead of linear. Hope this helps....

3ヶ月 前 | 0

| 採用済み

回答済み
How to train and test time series data in Neural Network Toolbox
I order to test a net you have to compare the actual output with the desired output. Hope this helps. THANK YOU FOR FORMALLY...

3ヶ月 前 | 0

回答済み
Artificial Neural Network implementation and to know the importance of each of the input on output(Response) - wanted help
The way I determine the importance of a single input is 1. Calculate the error using all inputs 2. Loop over inputs ...

3ヶ月 前 | 0

| 採用済み

回答済み
transfer function purelin equation based on neural network toolbox
Read the documentation help purelin doc purelin Greg

3ヶ月 前 | 0

回答済み
How to use a sequenceInputLayer with a regressionLayer (neural networks) ?
The answer is obvious: help regressionlayer doc regressionlayer *Thank you for formally accep...

4ヶ月 前 | 0

回答済み
Time Domain Signal for neural network
If you have rpm I don't see why time is important. However, there is no reason you cannot do both and see what difference it ma...

4ヶ月 前 | 0

| 採用済み

回答済み
Adding hidden layers to a patternnet hurts accuracy?
The global minimum is achievable with a single hidden layer. With more hidden layers you add more local minima; most of which ...

4ヶ月 前 | 0

| 採用済み

回答済み
Training neural network - Forward Pass
You are very mixed up. I suggest going to the library and finding a good elementary NN book. Obviously, the appropriate info is...

4ヶ月 前 | 0

回答済み
Hello, I'm working with artificial neural network.
Three layers are sufficient: input/hidden/output The input layer is NOT a neuron layer. The number of input nodes is the dimens...

4ヶ月 前 | 0

| 採用済み

回答済み
Why neural network gives negative output ?
How different is the new data (e.g., Mahalanobis distance)? If you know the true outputs, how do the error rates compare? If y...

4ヶ月 前 | 0

| 採用済み

回答済み
Rsq from NMSE in NN
NMSE = mse(trnopdb-net(trnipdb))/MSE00 i.e., NO TRANSPOSES 2. Rsq = R^2 3. Yes. Use separate calculations fo...

4ヶ月 前 | 1

| 採用済み

回答済み
HOW DO I SEARCH IN ANSWERS ???
Oh! It only opens up at the top of the page if the page is sufficiently wide. Since I often use large type and 2 pages per scre...

4ヶ月 前 | 0

| 採用済み

回答済み
Interpolation using Neural Networks
Plot the data. Look at the plots Are there regions that are seasonal? Separate the data that is relevant to your problem Ca...

4ヶ月 前 | 0

| 採用済み

回答済み
How to plot only few classes for confusion matrix?
Oviously, you have to subsample the original matrices. Greg

4ヶ月 前 | 0

回答済み
Open loop Training performance and closed loop training performance are good but multi-step prediction is bad. Reason?
You are not considering information from the autocorrelation fuction. Hope this helps Greg

4ヶ月 前 | 0

回答済み
Formula for two layer FFNN
y1 = b1 + IW1 * x y2 = b2 + LW2 * tanh( y1 ) y3 = b3 + LW3 * tanh( y2 ) = b3 + LW3 * tanh( b2 + LW2 * tanh( b1 + IW1 * x ...

4ヶ月 前 | 0

| 採用済み

Load more