Penerapan Metode Exponential Smoothing Pada Prediksi Dana Donatur Di Lembaga Amil Zakat Ummul Quro Kabupaten Jombang

Abstract

Donor Fund Prediction is a prediction system that aims to predict donor funds at the Ummul Quro Amil Zakat Institution, Jombang Regency.  The Donor Fund Prediction is used to predict the next year based on previous year's data. This study focuses on using exponential smoothing while for the prediction method initially using moving averages.  The calculated data is 2014-2018 while 2019 is used for testing prediction errors.  In predicting exponential smoothing, alpha constant value which has the smallest error is needed.  To get it, several stages are needed, namely, predicting 2014-2019 using a moving average with a constant defined by the user, predicting 2014-2018 with exponential smoothing with an alpha value between 0 to 1, looking for the MAPE (Mean Absolute Percentage Error) value at each value  alpha used.  After obtaining alpha with the smallest MAPE, the alpha value is used to predict 2019. The test results explain that calculations using a program with the Moving average method and Exponential smoothing successfully predict with an accuracy of 93.32% or only have an error of 6.68% instead of using only the method.  The moving average only has an accuracy of 90.25%.

Keywords: Prediction, Exponential smoothing, Moving average