If correlation is linear, You may use multiple linear regression.

http://www.mathworks.de/help/stats/multiple-linear-regression-1.html

http://www.mathworks.de/help/stats/linear-regression-model-workflow.html

if you know that your equation is y=a*x1+b/x2+x3/c;

you can easily use PSO optimization to find the parameters a,b,c.

I can send you the Matlab code, if you send me your data.

Good Luck

It is multiple linear regression. Pl be sure before going for the mentioned model, otherwise you may opt other form of model.

Any statistical packages, even excel can solve your problem. It is simple in doing.

Just plot the function y=f(x) and use then the analysis tab in Origin and find the fitting section (you have there linear, polynomial fitting). So defining a polynomial fitting of y=f(x) using a thrid or four degree, you will have your correlation parameters.

Thank you Hossein Soleimani, i will send my data but i need your email id. Can you please send me or type here, my email id is [email protected].

Dear Krishna, if you are haivng OriginPro8, you can go to Analysis -> Fitting -> Nonlinear curve fit -> Open dialogue -> Settings -> Category -> User defined -> new -> ...and define your function, by stating independent and dependent variables with the constants a, b and c. and call the column data of x1, x2 and x3 to the file and fit. You will have the results in your hand.

Dear Krishna, this is my email ID:

[email protected]

As already mentioned by Sree rama murthy, you can use Origin to build your own functions you want to use for fitting your data.

Just follow his instrucions to create your own fitting function. However if your data are more complex (say a peak function) you need a good starting point for your paramters otherwise you might have trouble finding the right values. Parameter initialisation strongly depends on the complexity of fitting function. If you can give me more information about your actual problem I might give you some tips.

Hope that helps,

Moritz

https://www.researchgate.net/publication/297245624_Particle_Swarm_Optimization_Algorithm_and_its_Codes_in_MATLAB

Research Particle Swarm Optimization: Algorithm and its Codes in MATLAB

https://www.researchgate.net/publication/296636431_Codes_in_MATLAB_for_Particle_Swarm_Optimization

Code Codes in MATLAB for Particle Swarm Optimization

https://www.researchgate.net/publication/280244004_A_comparative_study_of_metaheuristic_optimization_approaches_for_directional_overcurrent_relays_coordination

Article A comparative study of metaheuristic optimization approaches...

https://www.researchgate.net/publication/297245624_Particle_Swarm_Optimization_Algorithm_and_its_Codes_in_MATLAB

Research Particle Swarm Optimization: Algorithm and its Codes in MATLAB

https://www.researchgate.net/publication/296636431_Codes_in_MATLAB_for_Particle_Swarm_Optimization

Code Codes in MATLAB for Particle Swarm Optimization

Dr. Krishna Srihari Bonasi You'll have to fit your data with an appropriate function. I have worked a lot on the different kinds of fitting. In this video, I have discussed the multiple peak fitting. How to customize a fitting equation or how to generate one’s own fitting equation. I have discussed in detail all the dependent variables. In the case you want to further ask about it, please do comment on the specific video, I'll respond to it shortly. I have provided the practice file(s) here. Thanks

https://youtu.be/ViIdEllu5pA

While, for other fittings, one may see as, like linear fitting (https://youtu.be/ViIdEllu5pA), polynomial fitting (https://youtu.be/JrL5ebTq5qM), nonlinear curve fitting (https://youtu.be/hkjsvbo6U30), exponential decay fitting (https://youtu.be/pdiOc86IZ6U), histogram fitting (https://youtu.be/cfW88FZzsyA), and Nonlinear fitting with constraints (https://youtu.be/XLdxjW1iQmc).