Penggunaan Metode ARIMA untuk Memperkirakan Permintaan Obat-obat yang Dikelompokkan (Clustered) Berdasarkan Turnover Persediaan

Abstract

Pharmacy X is a drug store that sells variety of drugs that have several pharmacy branches and one central pharmacy. However, each branch and center of the pharmacy have their own administration and server, which cause the central pharmacy is difficulty controlling. In addition, an increase and decrease in drug demand causes Pharmacy X difficulties in determining the right time to buy drugs. With this problem, we need a system that can predict sales for each drug in the future and a centralized system so that the central pharmacy can control the branch pharmacy. To predict the sales, one of the prediction methods is Autoregressive Integrated Moving Average (ARIMA). ARIMA is a method of predicting using time series data from the previous period to predict future periods. For making a centralized information system using Slim Framework with PHP version 7 and using a MySQL database. The display of the system uses the HTML programming language, CSS, and Bootstrap. The results testing of the ARIMA method performed for each drug using R-Squared obtained 1.87% good accuracy, 35.06% medium accuracy, and 63.07% poor accuracy. Good accuracy is 1-0.5, accuracy is 0.5-0, and bad accuracy <0. For information systems the results are in accordance with the needs of the problem controlling the central pharmacy to the branch pharmacy. The synchronization feature makes it easy for central pharmacies to monitor sales and stock of goods at branch pharmacies.