A system identification methodology that accounts for factors influencing the reliability of identification has been proposed in previous work. An important aspect of this methodology is the generation of a population of candidate models. An indication of the reliability of system identification can be obtained through an examination of characteristics of the population. This paper presents feature extraction techniques that provide support for this examination. We use data mining techniques such as correlation measurements, principal components analysis and decision trees. These methods are useful for estimating the reliability of system identification when the number of candidate models is high and when relationships between parameters can be approximated linearly.