Thank you for this great work.
My question relates to the method linearinterpolate in file utils/manipulator.py.
In the following line of code:
> linspace = linspace - latentcode.dot(boundary.T)
I suppose you intend to compute the distance from the SVM boundary line using the term _latentcode.dot(boundary.T). However, this term computes the distance only if the boundary line passes through the axes origin. That is, if the constant in the decision function (i.e., boundary line function) is zero.
Can you please let me know whether I am missing something?
Thank you for this great work. My question relates to the method linearinterpolate in file utils/manipulator.py. In the following line of code: > linspace = linspace - latentcode.dot(boundary.T) I suppose you intend to compute the distance from the SVM boundary line using the term _latentcode.dot(boundary.T). However, this term computes the distance only if the boundary line passes through the axes origin. That is, if the constant in the decision function (i.e., boundary line function) is zero. Can you please let me know whether I am missing something?