Error Reduction of Parametric Auto Regressive Moving Average Estimation Algorithm to Reach Optimal Point

There are several methods for parameter identification and estimation in linear and nonlinear multi-input and multioutput systems. The method of detecting and identifying system parameters in a simulated auto correlated system with a minimized error between the target data and output results is affected by noise. After adding the summed noise to the input data, the system coefficients are determined in the least-squares algorithm. The trend of system error changes decreases with increasing number of input samples, but this does not mean that the estimation error decreases with increasing iteration with increasing trend. In other words, the number of samples given to the system in order to obtain a specific error does not decrease with increasing number of replications. Also, reducing the estimation error of the system is more dependent on input data than on data replication.