Try Anfis. researches indicated to the point that Anfis results are better than ANN

Can you give more information about the dimensionality of your data. Also when you tried SVM did you do  a grid search? or tried different kernels?

Hi Achraf 

My training data is 6D because I used 6 seismic attributes. I ran the SVM using Gaussian kernel.

You can go with ANFIS. 

Well you will be surprised what happens when you vary the parameters of the ref kernel. You might also try different kernels. 

Hi Samer,

I tested DT and it is not appropriate. It is time-consuming for this data and its result is not good. 

Thanks,

Amir

did you try decision tree?

Hi Noura,

Yes, I used this algorithm.

Sorry for now being a little bit straightforward. It is not the question with which classifer method you have play around. First you should answer yourself the question what are your features and are they able to discriminate in general? Is it worthwhile to apply a feature reduction / feature space reduction? What about the generalisation ability of your classifier? Robustness in context of your application? Measures of quality? Receiver operation curve? etc. 

Iff you know your features then you are able to decide which potential classifier you might use. But this is the second step. I also do recommend a book to my students: https://mitpress.mit.edu/books/introduction-machine-learning.

Godd luck!

Dear Amir,

I agree with Volker - the first step is data preprocessing (defining good features and also good training samples). I do not now this is classification or regression problem. In the case of classification problem lets start using the simplest classifiers such as k-nearest neighbors or naive Bayes. In many cases they give best results.

Regards 

Use ensemble of classifiers. Ex. Ensemble of svm with different kernels shows high performance.