Yeah I missed the part where you were testing on slices. Very nice. How long did it take you to come up with this workflow?
It took me a couple of years to write the software. The workflow is not complicated, just writing the software to do it was time consuming. I am launching my product in a couple of weeks and the site will have an extensive video collection showing exactly what I am doing.
There is a video on the first page of my journal. I will have some shorter ones when I launch. https://www.elitetrader.com/et/threads/adventures-in-automation.329052/
%% Yes;AlphA, i use machines that learned how to draw 200 day moving average. But best to use discretion also; 4 chart services have 4 different values for 200 dma . Actually many think MAs do not work, LOL well OK,then, you do the work.With the track record of most HFs, ignore about 80%, or 95%..............................................................................................
I outlined my current machine learning methods in https://www.elitetrader.com/et/thre...your-edge-for-2019.329802/page-8#post-4809209 and https://www.elitetrader.com/et/thre...your-edge-for-2019.329802/page-8#post-4809655
Yeah, mainly I was thinking like giving the machine bunch of market features and it tells me what is random and what is not, and what has happened before or not. I am wondering how trustworthy that would be considering issues such as overfitting or just imperfect training data. But I agree human judgement is needed.