don't know if anyone can help me here.

....

3 commentaires

Walter Roberson
Walter Roberson le 20 Nov 2013
What kind of analysis do you want to do? What is it about the array that you want to analyze?
Marc
Marc le 21 Nov 2013
How about nlinfit from the stats toolbox. If your heart is set on neural networks, start reading the documentation.
Otherwise, show us some attempt at some work and maybe we can better help.
Image Analyst
Image Analyst le 23 Nov 2013
Your subject line (at least at the current moment) says "don't know if anyone can help me here". I'd say you're making that very very unlikely with your scorched earth policy of deleting your past questions. I probably won't be helping you anymore.

Connectez-vous pour commenter.

Réponses (1)

Image Analyst
Image Analyst le 20 Nov 2013

0 votes

Try griddata(), TriScatteredInterp(), or scatteredInterpolant() depending on what version of MATLAB you have. Or if you have an image and need to fill in "holes" due to missing/black pixels in the image then try roifill() in the Image Processing Toolbox.

10 commentaires

Shani
Shani le 20 Nov 2013
I think I will use the neural network toolbox. I haven't got the slightest how to use it. any idea?
Image Analyst
Image Analyst le 20 Nov 2013
I don't have it either. Why do you think that NNs are the best way to get the values that you don't have or are missing?
Oh, wait a minute, when you say "values you don't have" you're not talking about missing values, like your data stream dropped out for a few seconds or your profilometer was unable to generate a height data value for that pixel or something like that. Perhaps you're talking about just simple interpolation or extrapolation via some model or equation. So you train it (build the model/equation) with some values and you want estimated values for some other values. Is that it?
Shani
Shani le 20 Nov 2013
Yeah basically, I have the weighting of a set of data series, which weighs the relative contributions of the principal components of the data set. Using this sets of weights/ information, the relating parameters are to be used, so I can predict values not originally in the data series. and I am a basic user for matlab I would not be able to build a model or non linear equation as I have basic understanding unless someone majorly helps me
Image Analyst
Image Analyst le 20 Nov 2013
I don't know what you have right now. Only data? Or were you able to get model weights by using something like polyfit()? And you need to know how to get estimated values (at locations not included in the training set you used to build the model) using polyval()?
Shani
Shani le 21 Nov 2013
I have got the model weightings, not using polyfit. Just simple mathematics. I am not sure how to use polyval but yeah I need to get predicted values. I need to create a neural network toolbox tho.
Image Analyst
Image Analyst le 21 Nov 2013
Then why can't you just do y = a1 + a2*x + a3*x^2 + .....?
Shani
Shani le 21 Nov 2013
Because I need to use the neural network for prediction. There is no one to help on this issue, I have tried on matlab answers, do you know anywhere where I could get help using the neural toolbox? like another link or something? please
Image Analyst
Image Analyst le 21 Nov 2013
Greg Heath is the NN expert here. You mgiht just have to wait until he logs in. Did you list the Neural Network Toolbox as a product? No, it looks like not. So that may be why he did not open your question or see it. Why don't you do that? Or else call the Mathworks? Do you have a full license with maintenance contract?
Zeeshan
Zeeshan le 21 Nov 2013
or can I use my university to contact them?
Image Analyst
Image Analyst le 22 Nov 2013
The student version does not qualify for support (except for installation support).

Connectez-vous pour commenter.

Catégories

Question posée :

le 19 Nov 2013

Commenté :

le 23 Nov 2013

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by