Peramalan Permintaan Produk Fashion Di PT. Behaestex Menggunakan Metode ARIMA Box-Jenkins

Fluctuating demand for products is a challenge for manufacturing companies, including PT. Behaestex is a well-known textile manufacturer in Indonesia. This study aims to predict the amount of demand for ATLAS Favorite brand sarongs using the ARIMA (Autoregressive Integrated Moving Average) method as an alternative to conventional methods that were previously based on subjective opinions. The sales data used are monthly data in the Hijri calendar for the period from Jumadil Awal 1441H to Rabiul Akhir 1446H. The analysis process is carried out through data stationarity identification, model parameter estimation, significance testing, and testing of the residual white noise assumptions and normal distribution. The results of the study show that the ARIMA model ([12],0,[1]) is the best model with a MAPE value of 56%, MAE of 6483.267, and MSE of 87447422.8 which is the lowest value and meets all statistical assumptions. This model is used to forecast demand until the Jumadil Awal 1446H period with a total order of 186019.6 dozen within one Hijri year. These findings are expected to be the basis for more accurate production and procurement planning at PT. Behaestex.