NonLinear SMV about tensorflow_cookbook

[ozgurgundogan]

Thanks for your response. Actually I got my answer from your notebook. Thanks a lot again :) You can close the issue.

[nfmcclure]

Hi @ozgurgundogan, thanks for bringing this to my attention. There was some extra code lines in there from when I was exploring how to write this code. I cleaned up the code a bit. But I also added my best quick explanation in an Jupyter notebook here: https://github.com/nfmcclure/tensorflow_cookbook/blob/master/04_Support_Vector_Machines/05_Implementing_Nonlinear_SVMs/05_nonlinear_svm.ipynb

For better explanations, I would recommend reading these links:
http://cs229.stanford.edu/notes/cs229-notes3.pdf
http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf

As a note, there needs to be the x and y points in any prediction you do because it's the x and y points that create the SVM margins. If you do some exploring with pred_kernel and my_kernel variable outputs, you'll see which non-zero values in the arrays indicate which data points are the support vectors.

I'll leave this issue open for a while if you want more clarification, otherwise I'll close it in about 48 hours. Thanks!