Predicting The EUR/USD Exchange Rate: A Deep Learning Approach
Authors
Nadj, Jonas Lund ; Rønsholdt, Asger Bo ; Jensen, Daniel Labuschagne
Term
4. term
Education
Publication year
2024
Submitted on
2024-06-03
Pages
117
Abstract
This thesis uses Long Short-Term Memory (LSTM) neural networks—a type of artificial intelligence that learns from sequences—to predict the direction of movements in the foreign exchange (FOREX) market. We combine macroeconomic data and technical indicators with several LSTM model variations to explore how these inputs can inform daily forecasts. The work is presented as a starting point for future, more complex models that may improve day-to-day predictions in FOREX and other financial markets. In addition, the thesis outlines the main building blocks of deep neural networks to make the approach easier to understand.
Dette speciale bruger Long Short-Term Memory (LSTM) neurale netværk—en type kunstig intelligens, der lærer af sekvenser—til at forudsige retningen for bevægelser på valutamarkedet (FOREX). Vi kombinerer makroøkonomiske data og tekniske indikatorer med flere varianter af LSTM-modeller for at undersøge, hvordan disse input kan informere daglige forudsigelser. Arbejdet præsenteres som et udgangspunkt for fremtidige, mere komplekse modeller, der kan forbedre dag-til-dag-forudsigelser i FOREX og andre finansielle markeder. Derudover beskriver specialet de vigtigste byggesten i dybe neurale netværk for at gøre tilgangen lettere at forstå.
[This apstract has been rewritten with the help of AI based on the project's original abstract]