Hlavní stránka Info portálu ZČU

What kind of data it is and how do you plan to use it in your research?
Eurex is an international electronic exchange belonging to Deutsche Börse AG group which primarily offers trading in European based derivatives. It is the largest European futures and options market. Eurex Enhanced Order Book Interface (EOBI) provides complete trade data and order book information showing every visible order and quote for the most liquid Eurex futures contracts. The EOBI data will be primarily used in research activities oriented towards the study of the so called rough fractional stochastic volatility (RFSV) models.

Could you explain to a general public, what kind of models you talk about?
In finance, volatility is the degree of variation of a trading price series over time. Stochastic volatility models are those in which the variance of the price process is itself randomly distributed, we say that the volatility is itself a stochastic process. We consider a new generation of stochastic volatility models, called rough volatility models, where the instantaneous volatility is driven by the so called rough fractional Brownian motion. This rough fractional driver is a short-memory generalization of the standard Brownian motion. Although RFSV models have proven to be superior in low-frequency market data fitting, very little is known about the roughness of volatility in the high-frequency intra-day data and about its evolution in big order book data over longer periods of time. Moreover, techniques used in the calibration of rough volatility models are still inefficient in terms of speed, accuracy, robustness, and sensitivity to be used widely in practice.

What does "big data" mean in your case? How big is the data set?
Well, we speak about billions of order book messages per day. Raw data has several hundreds of gigabytes per day and one-year data set is in the order of several tens to hundreds of terabytes. Roughly the same capacity will be also needed to decode and process the data.

That is really a lot of data. Where and how do you plan to process the data and what do you plan to get out of it?
We plan to use the computing and storage facilities of the Czech national e-infrastructure for research and development e-INFRA CZ, in particular the highly-intensive computations will be run in the MetaCentrum grid infrastructure that we are part of. Non trivial amount of time will be spent on understanding the data and reconstruction of all studied order books, i.e. it is hard to predict the results now, but we have several ideas and I can probably say that we aim to supply the academic and financial sector with the right accurate tools and techniques together with easily tractable algorithms. Among others, we will try to apply e.g. modern attention mechanisms from machine learning to develop innovative, highly accurate models based on deep neural networks. Here I would like to appreciate the help of my co-author Kamil Ekštein from the Department of Computer Science and Engineering.

Do you plan to involve students in your research?
Yes, of course. Suitable subsets of the extensive EOBI data set will be used for teaching purposes of my graduate level course KMA/SF Stochastic Finance, as well as for the individual student projects, including the Bachelor's, Master's and Doctoral theses.
Last but not least, I would like to take advantage of this opportunity and thank Deutsche Börse for giving us access to their data. I strongly believe that real market data will significantly enhance the quality of our academic and research activities.

Photos courtesy of Deutsche Börse AG.