This case study delves into an in-depth analysis of the Singapore Dollar (SGD) exchange rate trends against the United States Dollar (USD) from 2000 to 2019, using data sourced from the U.S. Federal Reserve's Download Data Program. The reader can expect the following insights and methodologies applied throughout the study:

The case study begins with a detailed decomposition of the SGD exchange rate time series to identify underlying components such as trend, seasonality, and irregular fluctuations. An analysis of these components provides foundational insights into the behavior and characteristics of the SGD exchange rate over time.

Based on the decomposition and analysis, a suitable time series forecasting algorithm is selected for predicting future values of the SGD exchange rate. Different time series models, including the recommended algorithm, are built and applied to forecast the exchange rate.

The 2019 data serves as a test set to evaluate the performance of constructed models using a rolling one-week evaluation window. The study compares the models' performances using chosen evaluation metrics, highlighting the most effective forecasting model for the SGD exchange rate.

Recognizing the potential influence of external factors, the case study explores the impact of the Chinese Yuan (CNY) exchange rate as an exogenous variable in forecasting the SGD exchange rate. A multivariate time series model is built to incorporate this external factor, with evaluation results and analyses provided to understand its effect.

The study concludes with considerations for operationalizing the model, focusing on strategies to ensure continuous performance and adaptability to new data and changing market conditions. Through this case study, readers will gain comprehensive insights into the methodologies and analytical techniques used in time series forecasting, particularly in the context of currency exchange rates. The findings not only reveal trends and forecasting accuracies but also shed light on the complexities of incorporating external factors into predictive models.