Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Master's Thesis |
| Scope | Discipline-based scholarship |
| Title | Optimizing Algorithmic Trading Strategies Through Reinforcement Learning |
| Organization Unit | |
| Authors |
|
| Supervisors |
|
| Language |
|
| Institution | University of Zurich |
| Faculty | Faculty of Business, Economics and Informatics |
| Date | 2021 |
| Abstract Text | This thesis explores the application of the Advantage Actor Critic (A2C) Deep Reinforcement Learn- ing Algorithm to a trading portfolio comprised of the 30 Dow Jones Index shares. Daily OHLCV data is downloaded from Yahoo!Finance and this price and volume data is supplemented with four technical indicators which, along with return and covariance data, constitute the features of the trading environment. The application is modelled as a Markov Decision Process, where the A2C algorithm is the agent acting upon the trading environment. The trading agent is trained with a deep neural network. The algo-driven portfolio return and Sharpe ratio are compared with the return and Sharpe ratio of the Dow Jones Index. Backtests are done for the period from 1 September 2020 to 1 September 2021. However, since this was an unusually volatile period because of covid-19 and US election concerns, the backtest is repeated for the relatively less volatile period from 1 January 2019 to 1 January 2020. We nd that with transaction costs set at 0.1%, the algorithm delivers an expected alpha of 5 percent and 1 percent respectively. |
| Export |
BibTeX
|