Forecasting Plastic Exports from Bangladesh: A Time Series Analysis

Abstract

key trends, and employing statistical models. It utilizes three different statistical models to predict. The best fitting ARIMA(p,d,q) model and the ets(A, A, N) model identified using the 'pmdarima' and 'statsmodels" libraries in Python are employed for forecasting. The ARIMA(0,1,1) model is recognized as the best-fitting ARIMA model, while the ets(A, A, N) model emerged as the most appropriate exponential smoothing model. Remarkably, the ets(A, A, N) model projects a much steeper growth trend in Bangladesh's plastic exports than the other models, even over 13 years. In addition to these models, the ets(A, Ad, N) model is also used, which assumes a dampened trend that becomes horizontal in the long run. This model is found to be the best-fitting ETS model. Across all three forecasting methods, a significant increase is predicted in Bangladesh's plastic exports in the future. Therefore, the study emphasizes the importance of implementing supportive policies to capitalize on this untapped export potential, positioning Bangladesh's plastic industry for sustained growth in the global market.