Dear Mohammad Ali Shariati,

Technically speaking, you should use the so-called symbolic regression methods. More information are provided at the attached links.

http://link.springer.com/chapter/10.1007/978-3-319-22180-9_70

http://www.sciencedirect.com/science/article/pii/S0020025512003295

http://booksite.elsevier.com/9781558605107/errata/p1.pdf

https://books.google.com/books?hl=en&lr=&id=NkG7BQAAQBAJ&oi=fnd&pg=PR7&dq=gene+expression+programming&ots=YYotvDWjFX&sig=RDNmT1Ue_YIu8SuxiucaGzKwfoo#v=onepage&q=gene%20expression%20programming&f=false

I recommend trying different models if you have no references of previous works in you related field. These papers may be useful.

Good Luck!

Article A neural network approach to weibull distributed sea clutter...

Article Improved Shape Parameter Estimation in K Clutter with Neural...

http://www.tutorialspoint.com/artificial_intelligence/artificial_intelligence_neural_networks.htm

the above link helpful to you regarding basics of artificial neural networks.