Forecasting of raw material needed for plastic products based in income data using ARIMA method

Abstract

Forecasting is a process of predicting something future by doing calculations from previous data. In this case the authors will forecast the sale of plastic production by using ARIMA Box-Jenkins method for 2015 forecasting. The data used is the sales data of plastic factory production in Bandung from 2012 to 2014. This research will use ARIMA procedure in SAS that allows for identification, Estimation and forecasting of Time Series models. The measurement of the accuracy of forecasting results is done with the MAPE (Mean Absolute Percentage Error) value. Forecasting results conducted for 2015 using ARIMA (3.0, 2) on plastic product sales data for 2012 to 2014 resulted in a prediction accuracy rate of 74% for PP Trilene and 68% for PP Tintapro products.