Answers are grounded in this paper and cite their sources.
The identification techniques available today are first classified into three broad classes. Their behaviour with respect to selected criteria are then explored and assessed. Further, a series of simulations are carried out with various types of bad data. Investigating the way these identification techniques behave allows completing and validating the theoretical comparisons and conclusions.
No takes yet. Share an insight, caveat, or question.
Mili et al. (1985) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: