So I've been using ensemble methods to model stock price movement, using intraday per-minute data in the OHLCV format, with the prediction being a 1 if the future close goes up, and 0 if it goes down. There are rows within the data that have the same closing price, and due to my somewhat limited understanding of how to interpret intraday data, I do not know if this is a normal occurrence.

If I factor in that the stock price may not move at all, then this becomes a multiclass classification problem, as I then have 3 outcomes to consider. This obviously entails a more involved process in producing error metrics, so I wanted to know if I am mistaken in thinking this is a multiclass problem. Thanks

  • $\begingroup$ I think is not fruitful trying to model the next-minute closing price, or return for that matter. You'll need an impressive accuracy in order to overcome the transaction costs, unless you have DMA. Unfortunately, the above claim is mainly anectodical, so I can't point out any specifical paper. By the way, if you're still in for trying, you can always label your data with three different classes, and train your classification algorithm pairwise. $\endgroup$ – james42 Aug 1 '17 at 13:09
  • $\begingroup$ Thanks for the response. The best out-of-sample accuracy I was getting with my methods was around 70%, but this project is less for practical implementation and more for thought experiment anyway. What would you recommend as the price I should attempt to model, if I'm seeking to maximize return? $\endgroup$ – Hassan Sabree Aug 1 '17 at 14:00
  • $\begingroup$ A 70% accuracy is exceptional, so I would check that you didn't overfit the data, it happens often. The choice of the "target variable" is yours, basically: if you simply want to hold the position for a few candles, you use the closing price as a target variable, more complex strategies can be made by taking into account the fluctuations inside a candle. $\endgroup$ – james42 Aug 1 '17 at 14:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.