Answered
How to check and remove outliers when it is Non-normal distribution
Regardless of the distribution, I find that a combination of zscore with plots of original and transformed data is sufficient fo...

10 years ago | 2

Answered
to prove the robustness of neural network model what is the best model which can compare to it ? (especially in order to forecast)
No. I measure robustness by adding increasing levels of noise to the input. Hope this helps. Greg

10 years ago | 1

| accepted

Answered
Is it possible to define the result of the neural network inside the tanh (x), its value is between 1 and -1?
It is possible. However 1. It is not easy 2. There is no reason for doing so Hope this helps. *Thank you for forma...

10 years ago | 0

| accepted

Answered
Is possible to normalize just the training data instead of the whole data using Neural Network toolbox?
The normalization parameters ARE obtained from the training subset; then applied to all of the data. So, you have no problem!...

10 years ago | 0

| accepted

Answered
Forecasting electricity load based on various parameters using neural networks.
How may inputs? How many outputs? How much data? Typical time between measurements? Are measurements equally spaced ? ...

10 years ago | 0

| accepted

Answered
TDNN multistep prediction with unknown future data for the target
In general, PREPARETS will yield the correct inputs. However, for TIMEDELAYNET, just use common sense: Ai = {} % There is...

10 years ago | 0

| accepted

Answered
input transformation in backpropagation neural network (prediction task)
Regardless of the original target transformations, I suggest that you normalize MSE by the average target variance MSE00. The la...

10 years ago | 0

| accepted

Answered
Is possible to normalize just the training data instead of the whole data using Neural Network toolbox?
No. Why would you want to do it? It doesn't make any sense to me. Greg

10 years ago | 0

Answered
multi step early prediction in neural network
Search for my examples in both the NEWSGROUP and ANSWERS. Hope this helps *Thank you for formally accepting my answer* ...

10 years ago | 0

| accepted

Answered
When I generate a Matlab function of a neural network, it comes with the normalization/re-normalization constants. Will these constants affect the neural network performance on another data set? Why?
Not if both datasets have similar summary statistics (e.g., mean/std for mapstd or min/max for mapminmax). That is why someti...

10 years ago | 0

| accepted

Answered
By including a moving window of fixed length in the input vector of MLP, is the Back-propagation ANN equivalent to NAR model?
1. When you insert code try to make sure it runs. N= lenght(data); % ERROR d=timestep ahead; % ERROR 2. Replace TRAI...

10 years ago | 0

| accepted

Answered
Train more a Neural Network to have more consistent results every time?
Results are different because the initial state of the RNG is different. Therefore you will get different random data divisions ...

10 years ago | 0

| accepted

Answered
Problem with defining networks
I ran your code with no error messages. Therefore you need to check version, license and installment. Hope this helps. ...

10 years ago | 0

| accepted

Answered
Why is training set worse that validation and testing for pattern recognition network performance?
I have used your choice of [ H1 H2 ] = [5 4] with TRAINRP and obtained behavior similar to what you have described. Although ...

10 years ago | 1

| accepted

Answered
How to calculate confidence intervals with neural network prediction?
When my work has required prediction intervals I have just used [ y - sqrt(mse) , y + sqrt(mse) ] where y = net...

10 years ago | 1

| accepted

Answered
Spread of RBFNN for prediction
Use timeseries NARX for prediction. help narxnet doc narxnet Determine input delays via the significant lags of the inp...

10 years ago | 0

| accepted

Answered
How to see confusion matrix , training curve and ROC plot in newrb MATLAB?
1. NEWRB does not have a. Designer weight initialization b. Datadivision or validation stop options. c. Non-equal ...

10 years ago | 0

| accepted

Answered
how to show performance of RBF neural network?
It cannot be done because NEWRBE is created with all training vectors. It does not learn. You can try NEWRB which learns by ...

10 years ago | 0

| accepted

Answered
How to train RBF NN with MFCC coefficients??
UNFORTUNATELY, A given newrb net can be trained once and only once. Additional training is not allowed. Greg

10 years ago | 0

| accepted

Answered
NARX Tapped Delay Line
Consider this STATIC APPROXIMATOR to the OL (Open Loop) NARXNET: clear all, close all,clc, plt=0 [ INPUT TARGET ] = simpl...

10 years ago | 0

| accepted

Answered
HOW to use NN tool in MATLAB ? and then what are input to give?
You have to define the "I"nput matrix, x, with size I x N consisting of N I-dimensional input vectors and corresponding "O"utp...

10 years ago | 0

| accepted

Answered
Neural Network input error
input and target matrices should have the same number of columns [ I N ] = size(input) [ O N ] = size(target) Hope this...

10 years ago | 0

| accepted

Answered
neural network error Undefined function or variable 'network'?
It is clear that you do not understand how the MATLAB toolbox functions work. 0. MATLAB DOESN'T USE UPPER CASE. I AM USING I...

10 years ago | 0

| accepted

Answered
when to use mapstd, mapminmax
I prefer standardization for 2 reasons 1. Easier to detect outliers for removal or modification 2. Typically, 0<= MSE <=1 ...

10 years ago | 1

| accepted

Answered
How i indicate the weights and bias to begin neural network ?
SEARCH USING: fitnet tutorial NEWSGROUP 12 hits ANSWERS 22 hits Hope this helps. *Thank you for formally accep...

10 years ago | 0

| accepted

Answered
using perfcurve with neural network
Currently, I cannot help because I have never encountered this function. When I get time I will consult the documentation usin...

10 years ago | 0

| accepted

Answered
How many neurons in the hidden layer???
For a given performance on training data, the fewer the number of weights, the better. Then the net will better generalize to va...

10 years ago | 0

| accepted

Answered
Neural Network in loops: How can I set up a loop to train at least 10 neural networks with the same parameters and save only the best performance, regression and histogram error, and the matrix-only MATLAB function for neural network code?
Search the NEWSGROUP and ANSWERS using Hmin:dH:Hmax Ntrials for my double-loop design examples that are typically over ...

10 years ago | 0

| accepted

Answered
In the figure below the trained data (shown in green ) is not covering fully to the target data by Artificial neural network technique. I used feed forward ANN technique. Is it the limitation of the ANN or anything else
You have to decide how much of the target variance you want to model. For a regression net (e.g, FITNET) I try to model at least...

10 years ago | 0

| accepted

Answered
neural network sensitivity analysis inputs
No. However, I have gotten very good fast suboptimal results by starting with polynomial models of standardized (zero-mean/unit-...

10 years ago | 0

| accepted