Modeling and Forecasting of Time Series Data using Different Techniques
Shayla Naznin Arzo Ahmed · Zenodo (CERN European Organization for Nuclear Research) · 2021
Currency is an important economic indicator. The indicator has some significant impact on the development of a country. A few key models have been suggested to improve the accuracy and efficiency of modeling and predict time series. This study examines the effects of currency exchange rate under Random Walk Model, Single Exponential Smoothing, Double Exponential Smoothing and Holt-Winter Models and the forecasting performances of the models are judged by the measure of accuracy both symmetric and asymmetric loss functions are used Mean Square Error (MSE), Mean Absolute Deviation (MAD) and Mean Absolute Percent Error (MAPE). From the measure of accuracy, double exponential smoothing model can be used to predict and smoothing the series of currency exchange rate with the other three different model. By comparing several models with the smallest value of the Akaike Information Criterion (AIC), we assessed an Autoregressive Integrated Moving Average (ARIMA) model that could be used to predict the exchange rate of the South Asian Association for Regional Cooperation (SAARC).