Dear Mukesh:

When I have cuantitatve dates I categorize and then I analyze its. ButI never work with dichotomous data because with these only it is possible make analysis of frecuencies.

I hope my answer helps you,

Best regards and thank you for you interesting in our work

Consuelo

Thanks a lot, Maria.

Dear Mukesh Kumar

I have done some research in  this type of problem and I have already evaluated several algorithms. In my publications you can find the data description and the algoritms results.

I listed some references that can help you. Please feel free to contact me if you need any further information.

Good Luck!

Mary

It might be imposible to tell in advance which algorithm will give you the best model for your dataset. Finding the best algorithm and tuning its parameters is part of the work when you mine data. You can try a few different algorithms and for each of them do parameter tuning and eventually you will find the winner.

Thanks, Laci Mary Manhaes and Nikola S. Nikolov for your valuable suggestions.

Dear  Laci,

Thanks for your publication which you provide me for reference.

I really find this publication helpful for my work.

But please tell me in which sequence I read these publications because all these publications are related to each other.

Thanks

Mukesh Kumar 

since you are going to do "prediction", it means you have classification problem.

start with SMO, decision tree then KNN and see the difference.

Have you tried AutoWeka ?