Developed by Loki Zhu (LSE)
Exchangerature is a tool that focuses on short-term exchange rate prediction. The prediction target is the daily average price. The purpose is not to provide speculative suggestions, but to provide convenience for people abroad. The model currently focuses on the CNY/GBP rate only.
The project was independently developed by a student of Economics at the London School of Economics (LSE). All predictions are generated using daily-updated machine learning models trained on the past 10 years of data.
Whether you're managing international payments, planning travel, or simply tracking currency movements, I hope this tool brings you clarity—and perhaps a small sense of control—in a fluctuating world.
* CNY is treated as the domestic currency in all context in the project
* The website updates automatically every day at 2:00 AM (UTC+0).
This exchange rate forecasting system is powered by a Long Short-Term Memory (LSTM) neural network, which is a type of deep learning model specifically designed to learn from sequential time-series data. The model is trained using historical daily CNY/GBP exchange rate data and Macro indicators. The model focuses on daily average price predictions, which indicates general short-term trends.
The model adopts a two-layer neural architecture which shows the best performance during tests. This design enhances the model’s ability to recognize complex short-term dynamics while maintaining generalization.
Each day, the model analyzes the 12-day movements in the average exchange rate and predicts the average value for the next trading day.
All data are collected from a reliable public sources yfinance , ensuring transparency and reproducibility.
The model is tested with the previous 1 year's data. The graph shows the model's performance in the test set. In terms of overall performance, the model achieves:
96.69% trend accuracy
95.08% where the error is smaller than 0.05