A set-Membership Identification Algorithm for Time-Series Modeling Using ARMA Process

Abstract

This paper focuses on the identification of ARMA models for time-series analysis. The proposed approach is developed in the Unknown But Bounded Error framework. Two Ellipsoidal Outer Bounding (EOB) identification algorithms are proposed. The second algorithm is an extended version of the first one. The paper includes stability and convergence analysis. Some numerical simulations are conducted to validate the analysis and assess the effectiveness of the proposed algorithms.