Is changing the 'KernelScale' in the fitcsvm equivalent to changing the sigma value?

2 次查看（过去 30 天）
monkey_matlab
monkey_matlab 2016-10-25
评论： Akshar Agarwal 2017-6-23
Hello,
I am using this command:
cl3 = fitcsvm(X,Y,'KernelFunction','rbf', 'Standardize',true,'BoxConstraint',10,'ClassNames',[-1,1]);
to run the Gaussian Radial Bias kernel. I understand that the default value for sigma = 1. I was told that in order to change the sigma value, use this command:
cl3 = fitcsvm(X, Y, 'KernelFunction', 'rbf', 'KernelScale', 0.2, 'Standardize', true, 'BoxConstraint', 100, 'ClassNames', [-1,1]);
In this case, the sigma value can now be changed by modifying the 'KernelScale'. Is this a correct statement?
  3 个评论
显示 1更早的评论
Aniket Borkar
Aniket Borkar 2016-12-16
But where can we set the C and sigma value? Box constraint =100 and kernel scale =0.2. How does these settings can vary C and sigma?
Akshar Agarwal
Akshar Agarwal 2017-6-23
the C value is the box constraint. It is called box constraint because it sets bounds on the lagrange multipliers in the actual function.

请先登录，再进行评论。

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Statistics and Machine Learning Toolbox 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by