DESIGN OF AN INFORMATION SYSTEM FOR FORECASTING FABRIC DEMAND (CASE STUDY AT CAHAYA SANDANG)

DESIGN OF AN INFORMATION SYSTEM FOR FORECASTING FABRIC DEMAND (CASE STUDY AT CAHAYA SANDANG)

Moving Average, Exponential Smoothing, Linear Regression, ARIMA, LSTM

In the course of weekly fabric procurement from suppliers, Cahaya Sandang frequently encounters challenges in determining the optimal quantity of fabrics to be ordered. The uncertainty surrounding order quantities has led to a loss of consumer confidence for Cahaya Sandang. To address this issue, a research initiative was undertaken, focusing on the development of an information system for forecasting fabric sales at the Cahaya Sandang Store. The fabric sales forecasting system was designed using Python, incorporating various forecasting methodologies such as moving average, exponential smoothing, linear regression, Autoregressive Integrated Moving Average (ARIMA), and Long Short Memory Term (LSTM). These designated forecasting methods are exclusively applied to predict cloth sales for the subsequent week. The determination of forecasting methods was based on the Mean Absolute Percentage Error (MAPE) values derived from the existing forecasting models. The findings affirm the efficacy of the five employed methods, demonstrating a maximum error value below 10% for the prediction of Cahaya Sandang fabric sales. By implementing this fabric sales forecasting system, the decision-making process regarding fabric orders from suppliers is expected to become more streamlined and targeted.

Geraldy Setiawan1, Dedy Suryadi2

https://jurnal.stie-aas.ac.id/index.php/IJEBAR

2024-06-01

PERI IRAWAN

PDF

ENGLISH

TEXT