Forex Prediction through Financial Trend Analysis

Forecasting the exchange rates has been approached through several ways over the years as the exchange rates of a country holds very high importance in giving us an insight into the economy of a country which largely influences the trade and economic relations. We propose a solution to forecast the currency exchange rates by using a bi-directional Gated Recurrent Unit (GRU) model that learns the variations in different economic trends to forecast the currency exchange rates of Indian rupee when compared to a unit US dollar. The data for six different economic trend features from 1970 to 2020 of both India and USA were collected and used to train the forecasting model. The model was able to acheive an MSE value of 0.000704.