See FinalReportCOMP490.pdf for in depth description of the project.
Project completed under the supervision of Dr. Thomas Fevens. Use of LSTM-Neural Networks to predict the future values of the foreign exchange rates. Compared with other time series predictive models like ARIMA (Autoregressive Integrated Moving Average) and Exponential Moving Average. Used on 3 different data sets of exchange rates and 2 different time frames.
The data sets used: CAD/USD, USD/EURO & AUS/USD.
Time frames: 1 day and 5 days prediction.