In this study using a combination method or atau hybrid model Autoregressive Integrated Moving Average (ARIMA) dan Artificial Neural Network to predict drug stock so that it can help Total Life Clinic pharmacies to plan drug stock inventory. The data used is drug stock data from 2015 to 2019 at Total Life Clinic pharmacies in the form of a monthly drug stock time series. In the analysis process to validate the prediction results using Mean Absolute Percentage Error (MAPE), while to see the performance of the ANN using Mean Squared Error (MSE). the validation results have a small error with a MAPE value of 0.041503 on the drug Tofedex with an average predictive accuracy value of 99.95%. and also obtained a high error with a MAPE value of 14,049 with an average prediction accuracy of 85.95% on Ferospat Effervescent drug.
Copyrights © 2022