Model Structure Selection - An Update

• While the topic has a long history in research, model structure selection is still one of the more challenging problems in system identification. In this tutorial we focus on impulse response modelling, and link classical techniques such as hypothesis testing and information criteria (e.g. AIC) to recent model estimation approaches, including regularisation. We discuss the problem from minimum mean-square error and maximum-likelihood perspectives.

Ämnesord

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Reglerteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Control Engineering (hsv//eng)

Nyckelord

Impulse response

Impulse testing

Maximum likelihood

Classical techniques

Hypothesis testing

Information criterion

Minimum mean-square error

Model estimation

Model structure selection

Regularisation

Publikations- och innehållstyp

ref (ämneskategori)

kon (ämneskategori)