Regularized System Identification: Learning Dynamic Models from Data

It provides a comprehensive treatment of recent developments in kernel-based identification, leveraging the power of machine learning without losing sight of system-theoretical principles. It offers new insight on classical questions and paves the way to new and powerful algorithms for linear and nonlinear problems. Regression methods such as regularization networks and support vector machines are the basis of techniques that extend the function-estimation problem to the estimation of dynamic models.

Format: Paperback / softback
Length: 377 pages
Publication date: 14 May 2022
Publisher: Springer Nature Switzerland AG

This comprehensive open-access book offers a thorough exploration of recent advancements in kernel-based identification, particularly relevant to individuals seeking to learn dynamic systems from data. The reader is guided step-by-step toward a profound understanding of a novel paradigm that harnesses the power of machine learning while maintaining a firm grasp of system-theoretical principles in black-box identification. The authors' reinterpretation of the identification problem within the framework of regularization theory not only sheds new light on classical questions but also opens up avenues for developing innovative and effective algorithms for a wide range of linear and nonlinear problems. Regression techniques, including regularization networks and support vector machines, serve as the foundation for methodologies that extend the function-estimation challenge to the estimation of dynamic models. Numerous real-world examples further illustrate the comparative advantages of the new nonparametric approach in comparison to traditional parametric prediction error methods. The challenges addressed by this book lie at the intersection of multiple disciplines, making it of interest to a diverse range of researchers and practitioners in control systems, machine learning, statistics, and data science.

Weight: 617g
Dimension: 235 x 155 (mm)
ISBN-13: 9783030958626
Edition number: 1st ed. 2022