Deep Learning the Stock Market
Tal Perry

Thank you, Tal; I really enjoyed this, which came into my inbox through this morning, and I thought it extremely clear and well-articulated, even though I am slightly reeling at the moment in my AI learning from trying to come to terms with the multiple instances where backpropagation can fail or stall. I haven’t had enough time to study the code, but I do think that one of the most interesting aspects of all kinds of NNs and the analysis of data (Big or not) is the possibility that it may disclose things that we just don’t see under normal circumstances. Stock Markets may be unpredictable, or they may just be unpredictable by dummies like me. Whether there is a formula that can be printed on a t-shirt isn’t really the point: why should there be? Why even should the stock market or the universe be intelligible to human minds except in some hugely simplified form? Dimension-reduction using t-SNE and other algorithms tacitly acknowledges that the data-space is just too large and complex for human minds to grasp, but that doesn’t mean that there is no underlying process or system. So I look forward to getting to grips with this and to further developments. And thank you. John

Like what you read? Give John Puddefoot a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.