Identification of errors-in-variables model with observation outliers based on Minimum-Covariance-Determinant

Diagnostics subspace identification method of linear state‐space model with observation outliers

“…It should be noted that the MSE for Algorithm 1 is 0.2608, whereas it is 0.0577 for Algorithm 2. Furthermore, the MSE for [20] is 0.2584 and for [19] is 0.0509. Though, the MSE for [19] is less than MSE of Algorithm 2; however, the time required is almost five times that is required for Algorithm 2.…”

“…Furthermore, the MSE for [20] is 0.2584 and for [19] is 0.0509. Though, the MSE for [19] is less than MSE of Algorithm 2; however, the time required is almost five times that is required for Algorithm 2. The true and the estimated eigenvalues of the matrix A are displayed in Table 2.…”

“…It should be noted that the MSE for Algorithm 1 is 0.5072, whereas it is 0.1056 for Algorithm 2 and this shows that Algorithm 2 improves Algorithm 1. It should be noted that both [19, 20] cannot be applied in this case. Table 3 shows the true and the estimated eigenvalues of the matrix A .…”

“…We write the state‐space model as a linear regression model, then derive a diagnostic statistical test to detect outliers. The presented subspace system identification algorithm does not require iterative computations, so it is computationally less expensive than [19] and more efficient than [20]. There are some differences between diagnostic techniques for linear regression model, and those for state‐space model.…”