9
$\begingroup$

Introduction:
I have seen a plenty of articles/books regarding volatility forecasting applied to high frequency data, but none of them were dedicated to forecasting the actual prices (for example bid/ask of currency pairs, stock prices).
Example:
Forecasting Volatility using High Frequency Data (P. R. HANSEN, A. LUNDE)

I realize that forecasting hft data must be hard because of several problems:

  • How long model will be valid?
  • How to check such validity?
  • Will a new model be calculated on time? (NN and most of subtypes of NN need a lot of time to be recalculated)


Questions:
Given the above mentioned problems/restrictions I would like to ask you:

  1. What's the best way of forecasting high-frequency data?
  2. What kind of scientific papers/books should I follow in order to be up-to-date in this topic?

Personal feedback:
From my own research based on hft data I can write that generally only few models are better than naive forecasting (but only in some sub time series - few minutes etc.)

  • Kalman Filtering (calibrated accordingly)
  • ARIMA, ARMA models

Generally worse than naive forecasting are:

  • Holts linear model (ARIMA (0,2,2)
  • Linear Regression
  • Exponential Smoothing
  • Moving averages (different types)
$\endgroup$
  • 2
    $\begingroup$ why would anyone in his right mind tell you how to predict, say, stock prices? $\endgroup$ – LazyCat Sep 9 '15 at 15:58
  • $\begingroup$ I am not building a trading agent, basically I am doing a scientific research. I have shared my results and maybe someone will be willing to contribute to the topic and will share his thoughts and experience. $\endgroup$ – Robert Szóstakowski Sep 9 '15 at 19:28
  • 3
    $\begingroup$ It doesn't really matter what you do, a question "how to predict stock prices" = "how to make money". Your feedback is meaningless, unless, you specify how do you apply your forecasting techniques. E.g. I would suspect that linear regression doesn't work for you not because it's a bad tool, but because you apply in it a wrong setting. Finally, to the best of my knowledge, most of academic research assumes some martingale condition for stock prices, which implies no forecasting. If you are not in this setting, it may be useful to specify your assumptions. $\endgroup$ – LazyCat Sep 9 '15 at 20:42
  • 1
    $\begingroup$ I don't know about books but you can look at: market-microstructure.institutlouisbachelier.org/…. It's a a lots of conference on HFT mainly given by academics. You can look for the programme and google the name of the researcher, I think you can even download the powerpoints of the conference. the LPMA (french math laboratory) has a great team in HFT (Mathieu Rosebaum etc..) You look for them. $\endgroup$ – user3018144 Sep 11 '15 at 12:50
  • 1
    $\begingroup$ One recommendation: If you want to forecast high frequency data, you also need to take into account (and model) the strong periodicity behavior observed in intra-daily volatility (the "U Shape") , see Andersen, T. G., & Bollerslev, T. (1997). Intraday periodicity and volatility persistence in financial markets. Journal of Empirical Finance, 4(2-3), 115–158. doi.org/10.1016/S0927-5398(97)00004-2 $\endgroup$ – Malick Nov 26 '15 at 0:09
0
$\begingroup$

A good place to start would be to understand the environment in which HFT takes place. What causes prices to form on the lowest possible level?

What mechanisms do exchanges use to process orders, how would they affect the price discovery process?

This in generally known as market-microstructure, and there is a lot of literature available on the subject, but you should also augment that with just physical observation of the orderbook, trades and the resulting price movements to get a better understanding of the whole process.

Once you have an understanding of what you are dealing with, you can start to build and test hypothesis.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.