Skip to content

Latest commit

 

History

History
14 lines (10 loc) · 803 Bytes

README.md

File metadata and controls

14 lines (10 loc) · 803 Bytes

fx-machine-learning

NUS Investment Club Quantitative Research Machine Learning Project

This research project tries to replicate the research done by Plakandaras, Papadimitriou and Gogas (2015) on using the EEMD-MARS-SVR to predict the magnitude and direction of the change in fx currency pairs.

Data Downloading

Data was obtain from various sources

  1. Quandl
  2. Yahoo! Finance

Initial data downloading and compiling was done using the quandl_download.py and compile_indices.py.

Also, included is the data file downloaded from the 2 sources inside the data folder and a simple presentation on a linear regresion model to predict USD/JPY movements and the profit and loss from this trading strategy. (updated 4 November 2017)