svm in machine learning toolkit for regression

[ziedhosni]

Hi everyone,

I found the neural network for the regression in labview toolkit very useful. I found other methods like SVM and the kNN. They are mainly for classification as the labels are integers. Is there any way to use them for regression problems.

If not, is it worth to use mathscript to call the machine learning module in matlab, and after the generation of an executable , will it require matlab to be installed?

Cheers,

Zied

[luda.mattski]

Hi Zied,

I'm not exactly sure which toolkit that is, but if it was the LabVIEW Machine Learning Toolkit that was posted in the Community, it may be worth posting on the ongoing software thread to get the best answers.  As far as I can tell, there should be a way to do function regression using that toolkit (based upon its description) but I'm not sure how to implement it. 

Regarding your second question, it is possible to call that third-party code using LabVIEW if you have the Mathscript RT node.  If you are planning on running an executable, you should be able to use most of the functionality that you would normally find within LabVIEW's runtime engine.  Here is a list of functions that you cannot call from the LabVIEW runtime engine - i.e. from an executable - but can use natively within the development environment:

http://zone.ni.com/reference/en-XX/help/371361G-01/lvtextmath/_rt_suplvtextmath/

[ziedhosni]

Hi Luda,

I read all the post in the forum about the SVM but the answer is still confusing and it did not work for me. And yes I am talking about the LabVIEW Machine Learning Toolkit.

[alexderjuengere]

What are you trying to achieve i.e. what is the goal of your regression?

What entities do you want to fit?

For example, here somone used a SVM to do Color Classification/Regression in scrap metals:

https://lavag.org/topic/13082-support-vector-machine-classifier/

It predicts class membership based on associated variables. In my job, we want to classify scrap metal bits based on color (red green and blue), for example. Say we want to separate brass versus copper from a mixed pile. The Support Vector Machine is a "supervised" classifier meaning that we start with sets of known members of each class. Use those to "train" the classifier and create a model. The model defines a hyperplane in Red-Green-Blue space with brass on one side and copper on the other. (Technically, it could be more than one hyperplane, but for me, in practice, just one.)

[ziedhosni]

Hello,

Basically I am trying to compare the performance of the SVM model with the models generated with the artificial NN or the k-nearest neighbour. I would like to comapre the MSE and R2 of the 3 methods of a regression problem for a non linear system like in quantitative structure activity relationship.

Cheers,

Zied

[luda.mattski]

Hey Zied,

What steps have you taken within the toolkit to achieve this functionality so far?  Since this is a third-party toolkit, I do not directly support it and am not very familiar with its function.

Again, it may be more helpful for you to post on the thread where you downloaded the toolkit, since the users around that thread likely have more experience with specific functions of the toolkit.