International students struggle with fluctuating currency exchange rates affecting expenses like tuition fees and remittances. The project aims to develop machine learning models predicting exchange rates among key regions: US, India, Europe, UK, and Japan. Forecasting rates enables students to make informed financial decisions, managing expenses amidst currency fluctuations.
The dataset used in this currency exchange rate prediction project is from Yahoo Finance containing historical data for over 4 years (2020-present) on the countries' exchange rates. The "Close" column contains the target values that the ML model aims to predict.
- ARIMA: A time series forecasting model that captures linear relationships and temporal dependencies in the data.
- SARIMA: An extension of the ARIMA model that includes additional seasonal AR and MA terms to capture seasonal variations and trends.
- XGBoost: It combines multiple decision trees to create a robust predictive model using gradient boosting to optimize the performance of individual trees in each iteration.
- Convolutional Neural Network: In time series forecasting, CNNs can learn hierarchical representations of input sequences, capturing both short-term and long-term patterns.
ARIMA CNN Decision Tree Linear Regression