# Greg Heath

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

**Statistics**

RANK**25**

of 242,467

REPUTATION**6,608**

CONTRIBUTIONS

**33** Questions

**3,103** Answers

ANSWER ACCEPTANCE **
45.45%
**

VOTES RECEIVED**1,106**

RANK

of 16,990

REPUTATION**N/A**

AVERAGE RATING**0.00**

CONTRIBUTIONS**0** Files

DOWNLOADS **0**

ALL TIME DOWNLOADS**0**

CONTRIBUTIONS**0 Posts**

CONTRIBUTIONS**0** Public Channels

AVERAGE RATING

CONTRIBUTIONS**0** Highlights

AVERAGE NO. OF LIKES

**Content Feed**

Would this be considered underfitting?

A model is UNDERFIT if and only if No. of independent training equations < No. of unknowns Hope this helps Tha...

1 year ago | 0

Avoid exploding/vanishing gradient problem with NARX nets?

Use a higher open-loop peformance goal. Then lower the value after the loop is closed. It's been years since I've done this so ...

1 year ago | 0

How to make prediction from a trained NARX neural network?

You forgot to include the intial conditions: yz = nets(xz,xiz,aiz); Thank you for formally accepting my answer Greg

1 year ago | 0

Out of memory issue while training a Neural Network (NN), array exceeds maximum array size preference using backpropJacobianStatic

A single hidden layer is sufficient. Hope this helps Thank you for formally accepting my answer Greg

1 year ago | 0

Out of memory issue while training a Neural Network (NN), array exceeds maximum array size preference using backpropJacobianStatic

A single hidden layer is sufficient. Hope this helps Thank you for formally accepting my answer Greg

1 year ago | 0

Difference between function of sim and predict in Neural network?

help sim doc sim help predict doc predict Thank you for formally accepting my answer Greg

1 year ago | 0

Forecasting using GRNN (Generalized Regression Neural Network)?

Yes, Normalize input and output data to [-1,1]. Minimize the order for a sufficient polynomial fit. Minimize the number of ...

1 year ago | 0

CNN With unbalanced Data

Although you do not prefer data augmentation, duplication of the smaller dataset examples is probably the quickest and most reli...

1 year ago | 0

Question

How to find questions on neural neural networks.

I have been away for awhile: How can I list the neural network questions? Greg

1 year ago | 1 answer | 0

### 1

answerTrained neural network to optimize input variable

"OPTIMIZED INPUT" is ill-defined Just train another net x2 = f2(x1,y1,y2) Hope this helps, THANK YOU FOR FORMALLY ACCEP...

1 year ago | 0

Linear Integer Output from a Neural Network

The integer nature of the output DOES NOT HAVE TO BE CONSIDERED during training. It is sufficient to round the real valued ou...

1 year ago | 0

How to disable validation and test data set in neural network

You have to define net before modifying any properties. clear all, close all, clc [x,t] = iris_dataset; for i = 1:2 net ...

1 year ago | 1

Training data and Training target in Neural Networks

You cannot make any intelligent decisions until you have examined a plot of the data!!! (WRONG!!! Plotting the data first is ...

1 year ago | 0

| accepted

simulink neural network producing different outputs to workspace

A simpler solution is to ALWAYS begin the program with a resetting of the random number generator. For example, choose your favo...

1 year ago | 1

How to avoid getting negative values when training a neural network?

Use a sigmoid for the output layer. Hope this helps THANK YOU FOR FORMALLY ACCEPTING MY ANSWER GREG

1 year ago | 0

weird plotregression plots for a 10% of my fitnet neural networks

Sometimes training gets into a parameter space rut. That is why it is wise to train multiple models. Hope this helps. Greg ...

1 year ago | 0

What is the purpose of shuffling the validation set?

To impose and verify a consistent GENERALIZED path to convergence by avoiding repetitive anomalies. Hope this helps Greg

1 year ago | 0

How to force overfiting of Deep Learning Network for Classification

OVERFITTING = More training unknowns (e.g., weights) than training vectors. OVERTRAINING1 = Training an overfit network to...

1 year ago | 0

Neural Network Pattern Recognition

Targets are 1-dimensional unit vectors with 4 zeros and a single 1 . Thank you for formally accepting my answer. Greg

1 year ago | 0

| accepted

semanticseg producing marginally different values when inference is repeated

Use clear all, close all, clc, rng(0) on the 1st line Thank you for formally accepting my answer Greg

1 year ago | 0

combining two neural networks (net1 is trained & net2 is untrained) in one bigger network

Just 1. Save the outputs of net1 in a file 2. Use the file to train net2 Greg

1 year ago | 0

Problem with the TreeBagger Command

The sizes of the input function and output target must be [ I N ] = size(input) [ O N ] = size (target) Hope this helps...

1 year ago | 0

How do I create a neural network that will give multiple input and outputs?

ALWAYS arrange your data so that [ I N ] = size(input) [ O N ] = size(output) Hope this helps. Greg

1 year ago | 0

How to make a hybrid model (LSTM and Ensemble) in MATLAB

Replace your YES/NO data with either 1/0 or 1/-1. Hope this helps. Greg

1 year ago | 1

how to augment image data only for a specific class?

Separate class 0 and interpolate. If you have a good feel for the data you could extrapolate. However the latter might be tricky...

1 year ago | 0

NTSTOOL - How to get predicted values of "the future"?

The known time series is analyzed to yield a time-series model that uses past and present values to predict future values. The ...

2 years ago | 0

Timedelaynet output calculation principle

You did not include tHe 2 biases. Hope this helps. Greg THANK YOU FOR FORMALLY ACCEPTING MY ANSWER

2 years ago | 0

Understand number of weights of Neural Network

It is possible. In general, however, you don't have the slightest idea what choice would be significantly better than random. ...

2 years ago | 0

Using pca for features selections

PCA (Principal Coordinate Analysis) is a very useful method for regression (it ranks linear combinations of the original variabl...

2 years ago | 0

How do we decide the number of hiddenlayers in a PatternNet?

patternnet(10) indicates ONE HIDDEN LAYER WITH TEN NODES It is important to be mindful of the number of layers and nodes. The ...

2 years ago | 0

| accepted