Click here to Skip to main content
15,914,165 members
Please Sign up or sign in to vote.
0.00/5 (No votes)
See more:
I have implemented LIBSVM in eclipse for binary classification of genes (100 by 584 dataset). When I compute the weight vector I get values in the range -5.62429704166405 to 5.12529916688051. Similar classification has been done in a matlab code (using some dll file for the SVM classification, I do not have the source file so I am not sure what is there inside the dll files) and when I use the same dataset and compute the weight vector in matlab, I see values in the range - 0.00016487 to 0.0028375.
Number of misclassifications and support vectors is comparable though in both the codes.
It seems weird to find such difference in weight vector coefficients. I am using RBF and C-SVC and I keep trying different values of C and gamma. I don't know what model has been used in matlab. Does it mean there is something wrong in my code?

Thanks!
Posted
Comments
Member 10305598 28-Oct-13 22:35pm    
I re-checked the weight vector values using quadprog in matlab and the values are similar to the matlab code using dll files
Prdpsnh 8-Apr-15 14:03pm    
Hello sir i having same problem that how to find weight vector w in java imlementation of libsvm .can you please share code or guide how to do so thanks

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)



CodeProject, 20 Bay Street, 11th Floor Toronto, Ontario, Canada M5J 2N8 +1 (416) 849-8900