Perancangan Model Peramalan Permintaan Produk Teh Botol Kaca dengan Pendekatan Machine Learning pada PT Sinar Sosro

Abstract

An accurate sales forecasting is crucial to the profits earned because it affects the company's stock management. Machine Learning is an artificial intelligence that can reduce the risk of demand uncertainty by performing data recognition continuously and automatically with computational assistance. PT Sinar Sosro is one of the companies in industrial sector that produces various beverage products. The factory in Medan serves the demand from the provinces of Aceh, North Sumatra and West Sumatra. one of the products is the tea glass bottle. At PT Sinar Sosro, there are frequent differences between forecasting data and sales data, causing high error rates in production planning accuracy. This research is designed to analyse the trend of demand for glass bottle tea products, analyse the best model to predict future sales by comparing the accuracy value of the Machine Learning forecasting model with the existing forecasting model at PT. Sinar Sosro and design the use of the Machine Learning system so that it can be used. This research was conducted using the Recurrent Neural network (RNN) method as part of the Machine Learning approach. The data that was inputted to the programme was weekly demand data, calendar holiday data, temperature data, and population data. The forecasted data is weekly demand. Based on the company's historical data, a demand graph is obtained which has a cyclical pattern. From forecasting based on Machine Learning, the accuracy value is 99.47% with an error value of 0.53%, which means it is still below the tolerance limit given by the company. The forecasting results are then implemented into a website that displays the forecasting results visually so that users can easily understand the predictions made using the Machine Learning model.