2 ms·
The problem is: "70% accuracy" doesn't mean much, since he's framed it like a classification problem. It's more like a regression problem. Plus, the data used
by discardorama 12y ago
The problem is: "70% accuracy" doesn't mean much, since he's framed it like a classification problem. It's more like a regression problem.
Plus, the data used is only for a couple of days.
And finally: if someone had an ML model to reliably make money on the stock market, they wouldn't be writing about it; they'd be laughing all the way to the bank.
- dkfmn 12y agoThe last point is especially salient. Except there's no "if." Major firms are already doing this (and not talking about it).
- petegrif 12y agoMany years ago I was a trader. I recall watching analysts 'explaining' what had just happened in the markets. The question that naturally arose, was why they weren't speaking from their yachts.
- hayksaakian 12y agothe simple answer is that you don't have the same information in front of you that you did at the time. in hindsight, we can easily identify what is and is not relevant/influential, whereas in real-time anything could prove to be relevant/influential.