Research Review of Active Learning Based on Automata
In the process of object-oriented analysis,the basic behavior of real system is usually expressed by finite state ma-chine.Automata learning technology is a common tool for learning system behavior model.With the rapid development of automatic learning technology and the gradual deepening of research,this technology is increasingly widely used in practice.Large software systems often contain a large number of black box components,and their behavior model cannot be inferred manually.Therefore,ac-tive learning technology can be used to fully automatically learn the system or component behavior model.This paper investigates,arranges and summarizes the current learning framework,modeling forms and learning algorithms of automatic learning.The storage structure of the learning algorithm is classified,and the learning algorithm is analyzed according to the different ways of dealing with counterexamples.Finally,the challenges of automatic learning and the direction of future research are discussed.
active learningblack box assemblybehavior modellearning algorithmcounterexamples
国家科技期刊平台
NSTL
万方数据
