Forecasting stock prices using real-time limit order book information
keywords MACHINE LEARNING, OTTIMIZZAZIONE IN FINANZA, TRADING AUTOMATICO
Reference persons GIUSEPPE CARLO CALAFIORE
Research Groups Systems and Data Science - SDS
Thesis type SPERIMENTAZIONE E SVILUPPO
Description It is well known that the analysis of intra-day prices (or price/volume pairs) yields very limited success in predicting future trends of stock prices. However, with the advent of modern online trading platforms, a large amount of real-time data is available even to the common, non professional, user. Such data includes, for instance, information relative to five or even ten levels of the limit-order book, arriving at a relatively high frame rate (e.g., every 0.1 second). It has been recently argued in the literature that such a large amount of data may be exploited for the purpose of prediction of future price behavior. The objective of this thesis is to explore optimization-based and machine learning algorithms and techniques for extracting information from the order book data flow. The successful candidate is expected to have some basic knowledge of the functioning of financial markets, basic knowledge of optimization methods, a good command of Matlab and of the elements of Java programming. Part of the activity may be developed in partnership with a leading company involved in online trading.
Required skills Optimization methods, linear algebra, basics of financial markets, Matlab and Java programming.
Deadline 14/07/2016 PROPONI LA TUA CANDIDATURA