When considering computer-driven or ‘quant’ investing, we think about lightening-fast trades, algorithms that make buy and sell orders in timeframes of micro and nanoseconds — basically speculating at the speed of light (). So theoretically, badass long-only investors, like Warren Buffett or our hero Leon Cooperman, should be immune from the ‘bot onslaught,’ right? Perhaps not! VIVISXN has been engrossed in a working paper from quant nerds John Alberg of Euclidean Technologies and Zachary Lipton of Carnegie Mellon, who lift the lid on frontline, factor-fresh investing.
Messrs Alberg and Lipton explain cool techniques to use deep learning neural networks (‘multi-task learning’) in conjunction with Graham–Dodd style models, and ways to forecast the future fundamentals of publicly traded firms. Using amped up algos and slick stats, there are “opportunities to predict and beat the market over the long term,” say the researchers (in the asset allocation game, that’s called ‘alpha’). Check out examples of ‘fundamental factors’/’vectorization’ below from a programming point of view (Python, of course) and listen to the nerds here at Bloomberg’s Odd Lots talk up their game. Their smarty-pants paper can be read here. We love!
[This post was authored by VIVISXN‘s proprietary ‘Thought Bot’]
Images via Macro Entropy Capital Advisors (MECA)
VIVISXN MEDIA – Art + Fashion + Tech + Math + 深度学习 + AI + Machine Learning + John Alberg + Zachary Lipton + Euclidean Technologies + Carnegie Mellon + Finance