Forecasting Monthly Export of Readymade Garments by Removing Seasonal Impact

Various Bangladeshi authorities release time series data that takes seasonal effects into account. However, no adjusted series is offered. The developed world has completely different conditions since they broadcast seasonally adjusted series. Many seasonal adjustment techniques, such as classical and X-based techniques, are available; however they cannot be used in practice in accordance with Bangladesh's seasonal time series. So here we have executed X-11 and X-12-ARIMA which are known as X-based seasonal adjustment methods and some classical methods like SARIMA and MA to seasonal time series data collected from secondary source as economic trend revealed by Bangladesh Bank. We will use export of readymade garments which are monthly data and have seasonal impacts. The entire data collection must first be divided into training and test data. Next, the data was de-seasonalized, and future values were predicted using training and test data sets, respectively, to compute various forecasting errors such as MAPE, PMAD, MAD, and RMSE. We utilize several X-based approaches and traditional methods to compare the errors. We conclude that the seasonal adjustment strategy performs better and has fewer forecasting mistakes. In conclusion, we suggest the optimal seasonal adjustment technique for ready-to-wear exports and project certain future values based on that technique.