DETERMINATION OF THE BEST FORECASTING METHOD FROM MOVING AVERAGE, EXPONENTIAL SMOOTHING, LINEAR REGRESSION, CYCLIC, QUADRATIC, DECOMPOSITION AND ARTIFICIAL NEURAL NETWORK AT PACKAGING COMPANY

Lina Gozali (Universitas Tarumanagara)
Sharin Candra (Universitas Tarumanagara)
Andres Andres (Universitas Tarumanagara)
Natalia Velany Putri (Universitas Tarumanagara)
Frans Jusuf Daywin (Universitas Tarumanagara)
Carla Olyvia Doaly (Universitas Tarumanagara)
Vivi Triyanti (Universitas Tarumanagara)

Jurnal Ilmiah Teknik Industri
Vol 9, No 2 (2021): Jurnal Ilmiah Teknik Industri : Jurnal Keilmuan Teknik dan Manajemen Industri

PT. Peace Industrial Packaging is a company that produces Styrofoam and plastic bottles that are used as containers or places or commonly referred to as packaging used by other companies to place their finished products. After getting data on the number of requests and production results obtained every month then the data will be processed using several forecasting methods such as Single Moving Average; Double Moving Average; Weight Moving Average; Single Exponential Smoothing; Double Exponential Smoothing; Linear Regression; Quadratic Method, Method Cyclic; Decomposition Method; and Artificial Neural Network (ANN) Method. After conducting the research calculation, the following conclusions can be drawn. The right forecasting method used for the HBL 100 ML product is the ANN (Artificial Neural Network) method because it has the smallest error value, namely the MAD error method of 760.583554, the MSE error method of 863,032.834043, the SDE error method of 970.304264, the MAPE error method is 0.112530, and the MPE error method is 0.112530