Historical Data Segmentation for System Identification: A Comparison of Methods

  • Giulio Cesare Mastrocinque Santo University of Sao Paulo
  • Claudio Garcia University of Sao Paulo
Keywords: Condition number, Data mining, Dynamic response, Historical data, Numerical analysis, System identification


This article explores the challenging problem of scanning historical data in order to find useful intervals for system identification. A review of the main works related to the subject is presented and two methodologies are described in detail. A single algorithm structure is then presented based on the analyzed methods and practical examples are given, allowing one to understand how to apply the algorithm to massive data.