摘要：AbstractIn system identification, input selection is a challenging problem. Since less complex models are desireable, non-relevant inputs should be methodically and correctly discarded before or under the estimation process. In this paper we investigate an input selection extension in least-squares ARX estimation and show that better model estimates are achieved compared to the least-square ssolution, in particular, for short batches of estimation data.
关键词：KeywordsInput selectionSystem identificationARX-modelsARMAX-modelsSignal-to-noise ratio
