So I was gonna reply to the old thread
Re: Machine Learning and A.I. predictions for 2018
but i figured it was off the rails enough and old enough by now to warrant a new thread (and truly i had to re-read the beginning and didn’t even remember most of the thread at all!)
anyway I wanted to share some of my exploits. I don’t do much data mining development these days, the basics are kinda cracked and discoveries are far and few in-between. Though i probably should work some more on t-sne… a 3-4 year old project that sits idle for months and months at a time.
data mining is a lot like data compression if you will. there is only so much information to be gleamed from a data set. and most of the basics seem very well explored and understood. the trick is generally framing the question right. that’s not to say its solved, but there seem to be far less discoveries and new research these days.
A.I. is not to far behind but that’s a different beast (reinforcement learning that is).
Annnnyway the reason i decided to post here was to share how succesful i’ve been after all these years. (shameless plug inbound) I started integrating with collective2.com a few tears back (i mentioned it quandl and stocks in the other thread) and have gotten to a point where i’m moderately successful! (if you can call 30% gain in 6 months successful) now i dont know if it’ll last. i hope it does, but i might re-read this in 2 years and say “ah that was the beginning of the end.” or maybe, “ah those were simpler times. that fund doesnt even exist anymore” anyway, if you have a mind check it out. I’m pretty pleased with the results thus far.
https://collective2.com/details/132385285
note that, there are actually 2 other funds I have that are marked private I just transformed in to versions of this one. i’m still seeing if i can outperform this. mainly they are selecting from the same pool of target stocks, but either more picky, more greedy or making fewer/larger investments. regardless this one is doing pretty well right now :)
