Project

Back to overview

Self-Adapting Regime-Switching Models for Automated Financial Trading Systems

English title Self-Adapting Regime-Switching Models for Automated Financial Trading Systems
Applicant Maringer Dietmar
Number 138095
Funding scheme Project funding (Div. I-III)
Research institution Wirtschaftswissenschaftliche Fakultät Universität Basel
Institution of higher education University of Basel - BS
Main discipline Science of management
Start/End 01.03.2012 - 28.02.2014
Approved amount 100'378.00
Show all

Lay Summary (English)

Lead
Lay summary
This project has two main objectives: (i) providing a reliable method for continuous calibration of forecasting models for financial assets which are used for automated trading systems; and (ii) assessing how well such forecasting methods could perform in the Swiss market. The former problem will be approached with innovative search and optimization techniques. The latter will be tested using historical data in a traditional ceteris paribus analysis. Additional tests are planned using an artificial market to assess the market impact of these strategies and whether the identified strategies could have a potentially destabilizing effect on the market.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Transition Variable Selection for Regime Switching Recurrent Reinforcement Learning
Maringer Dietmar, Zhang Jin (2014), Transition Variable Selection for Regime Switching Recurrent Reinforcement Learning, in IEEE Computational Intelligence for Financial Engineering and Economics conference, London, UK.
Indicator Selection for Daily Equity Trading with Recurrent Reinforcement Learning
Zhang Jin, Maringer Dietmar (2013), Indicator Selection for Daily Equity Trading with Recurrent Reinforcement Learning, in 2013 Genetic and Evolutionary Computational Conference, Amsterdam, NL.
Two Parameter Update Scheemes for Recurrent Reinforcement Learning
Maringer Dietmar, Zhang Jin, Two Parameter Update Scheemes for Recurrent Reinforcement Learning, in IEEE World Congress on Computational Intelligence (IEEE WCCI), Beijing, China.

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
IEEE Computational Intelligence for Financial Engineering and Economics Talk given at a conference Transition variable selection for regime switching recurrent reinforcement learning 27.03.2014 London, Great Britain and Northern Ireland Maringer Dietmar; Zhang Jin;
Computational Finance and Econometrics and ERCIM Talk given at a conference Enhancing recurrent reinforcement by using regime-switching models 14.12.2013 London, Great Britain and Northern Ireland Zhang Jin; Maringer Dietmar;
2013 Genetic and Evolutionary Computation Conference Talk given at a conference Indicator selection for daily equity trading with recurrent reinforcement learning 06.07.2013 Amsterdam, Netherlands Zhang Jin;


Abstract

This project has two main objectives: (i) providing a reliable method for continuous calibration of forecasting models for financial assets which are used for automated trading systems with a special focus that can deal with alternative regimes and situations; and (ii) assessing how well such forecasting methods could perform in the Swiss market. The former problem will be approached with innovative search and optimization techniques. The latter will be tested using historical data in a traditional ceteris paribus analysis. Additional tests are planned using an artificial market to assess the market impact of these strategies and whether the identified strategies could have a potentially destabilizing effect on the market.
-