Comment from our dev @JerryYun44
: Due to the nature of PCA, even if the sign of the eigenvector solution is reversed, it remains the same solution. This part was missing in the code. For example, in the first training, the coefficient might be 0.3X1 – 0.1X2, and in the second
PCA Eigenvector Sign Ambiguity Issue in Code
By
–
