Trigonometric SVM Clustering Signal

I created a trigonometric SVM clustering signal utilizing k-means and finding the hypotenuse on fractional brownian motion samples simulating price data. The results do not account for taker effect on MM/toxic flow and scaleability has not been elucidated.

The Code:

The Results:

Play around with this one… I assure you, it’s simple; but has potential to be alpha with some tuning.