Hi Sanu,

SVM in MATLAB is easy. I have utilized the same for river discharge forecasting in my analysis.

I have attached a folder in the zip file (named svm2). Save it on your system. Set path for these folders and its sub-folders by going to File- set path- and then adding the folder and its sub-folders to MATLAB (This will facilitates definition for key functions in SVM operation). Then use the SVM code as given in the pdf in zip file (page 3-4).

Please note that x(trn) = training variables, y(tr_tr) = target,xtest(tin) = testing input variables in the code.

All the best.

Use LibSVM. It's easy to use. Google it and you will find it.

I think it is for classification problems.Can I use it for regression purpopse?

Hi Sanu,

SVM in MATLAB is easy. I have utilized the same for river discharge forecasting in my analysis.

I have attached a folder in the zip file (named svm2). Save it on your system. Set path for these folders and its sub-folders by going to File- set path- and then adding the folder and its sub-folders to MATLAB (This will facilitates definition for key functions in SVM operation). Then use the SVM code as given in the pdf in zip file (page 3-4).

Please note that x(trn) = training variables, y(tr_tr) = target,xtest(tin) = testing input variables in the code.

All the best.

@Vinit Sehgal: You have been of immense help. Thanks for sharing.

@ visit 

Hello during integration of libsvm with MATLAB . I am not able to find compiler. MATLAB only showing win 3. Compiler. But I already installed Microsoft c++ compiler in my laptop.

sorry @vinit

@vinit 

Input is -8760*3

output  is -8760*1

test input is-63*3

Following error is coming kindly suggest me-

Out of memory. Type HELP MEMORY for your options.

Error in svmkernel (line 72)

ps = zeros(n1,n); % produit scalaire

Error in svmreg (line 84)

ps=svmkernel(x,kernel,kerneloption,[],framematrix,vector,dual);

Error in kkk (line 12)

[xsup,ysup,w,w0]=svmreg(x,y,C,epsilon,kernel,kerneloption,lambda,verbose);

Seems like the kernal formed for your data is very big and hence is exceeding the heap memory allocation for your MATLAB.  You can follow the link for diagnostics: 

https://www.mathworks.com/matlabcentral/answers/153250-out-of-memory-help-needed

@Vinit Sehgal 

Thanku for your valuable suggestion. If you have any your paper on this SVM reasults . Kindly mail the paper on [email protected]

Sorry Harendra,

SVM isn't my specialty and so, wont be able to help you with good literature on this. I am sure you would be able to get some good literature on SVM online as the topic is well traversed.  

@vinit Thanku sir

Hello is it possible to get a code for bayesian optimization based support vector regression. i need to change other optimization techniques also.

Thanks

Thanks

clc; clear; close all; %% Load Data xall=xlsread('D:\paper-tree\svm\in.xlsx'); yall=xlsread('D:\paper-tree\svm\out.xlsx'); xtr=xlsread('D:\paper-tree\svm\intr.xlsx'); ytr=xlsread('D:\paper-tree\svm\outtr.xlsx'); xtst=xlsread('D:\paper-tree\svm\intst.xlsx'); ytst=xlsread('D:\paper-tree\svm\outtst.xlsx'); TrainInputs=xtr; TrainTargets=ytr; n=numel(TrainTargets); %% Design SVM C=200; svmstruct=svmtrain(TrainInputs,TrainTargets,... 'boxconstraint',C,... 'kernel_function','rbf',... 'rbf_sigma',24); nTrainData=480; TrainOutputs=svmclassify(svmstruct,TrainInputs); plot (TrainOutputs,TrainTargets); mseTrain = mse(TrainOutputs,TrainTargets)

Thanks