Pendugaan Parameter Model Dinamik dengan Metode Robust Median Absolute Deviation (MAD)

Abstract

Given a data set in a dynamical model with unknown parameter values, the most commonly used method for parameter estimation is the least square method. However, this method is very sensitive to outliers. Because of this, a robust method to handle outliers is used. There are several robust methods such as Least Absolute Deviation method and Median Absolute Deviation method. When the number of outliers is relatively small, the result of parameter estimation using Least Absolute Deviation method and Median Absolute Deviation method are relatively similar. However, when the number of outliers is relatively large, the Median Absolute Deviation method is the most robust method to outliers. The dynamical models used in this manuscript are logistic and SEI (Susceptible, Exposed, and Infectious) models with hypothetical data. The accuracy of parameter estimation was measured by the value of Mean Absolute Percentage Error (MAPE).