Improving nonlinear regression model estimation based on Coati Optimization Algorithm

The mathematical and social sciences together with engineering fields use Non-Linear Regression analysis as one of their primary techniques. Controls and modeling of Non-Linear systems rely heavily on parameters estimation as a crucial problem. This paper presents a brief examination of this issue and develops an effective COA algorithm for parameter estimation accuracy enhancement of six Non-Linear Regression models (Negative exponential model, Model based on logistics, Chwirut1 model , Hougen-Watson model, Dan Wood model , and Sigmoid model). Simulation tests showed that the Maximum Likelihood Estimation (MLE) method using the Coyote Optimization Algorithm (COA) achieved the best performance when selecting among different methods along with different samples sizes and the mean squared error criterion.