Industry: Finance

Area: FX Trading

Solution: Machine Learning

 

Background

Our client for this project is a large FX provider, covering all major currencies and markets. They also provide a wide range of bespoke Foreign Exchange Risk Management Solutions. For this reason, it is vital they are receiving real time, precise information on the fluctuation of these currencies.

The Problem

.The problem they faced was the inability to accurately forecast how currencies will perform in the near future. There is no way to make a definitive prediction about how a currency will behave, but it is possible to make well-informed estimates.

The Challenge

We were tasked with developing a Predictive Machine Learning System which could use historical FX data to make predictions on the fluctuation of a currency. In order to do this we had to gain a thorough understanding of the clients data structure, and develop architecture for aggregations and relevant data sources. We used these data sets to uncover patterns and trends, enabling us identify key factors influencing the value of a given currency.

The Results

After using large quantities of training data and making incremental changes, we were able to successfully develop and algorithm which significantly improved the accuracy of our clients FX forecasting. As the algorithm accumulates more data, it’s forecasts will become increasingly reliable.