The Decomposition Of Complex Concepts Based On AFS Theory And Its Application

In the modern life, complex concepts are not easy to be recognized by machines, especially some complex concepts are judged by intuition of human, some complex concepts are not easily to be recognized even by human themselves. So based on AFS theory, and by studying the way human recognize the complex concepts, and the comparation and the sort of the samples of the complex concepts, we expect to find the rules between complex concepts and simple concepts. After that, the machine can study and recognize the rules in order to make the machine to think and judge like human such that machine can take the place of human in some particular situation. Secondly, the diagnosis and recognization of complex faults is particularly difficult because of the mass data and the difficult relationships between faults. So the introduction of artificial intelligence can lighten the workload and task of human such as machine learning, pattern recognition and so on. There was discussion about the diagnosis of complex faults based on AFS theory and its application in [1]. The paper applies the decomposition of complex comcepts based on AFS theory into the diagnosis of complex faults which is different from the method proposed in [1].AFS Theory [2] (Axiomatic Fuzzy Set) is a new analysis method of fuzzy mathematic. The Sets and Lattic Theory in AFS theory is one of common analysis ways that moden mathematic uses. In AFS theory, there is relationship between data and attribute. Based on the relationship, Prof. Liu proposed a way about machine learning and artifical intelligence, which is widly used in pattern recognition, machine learning, artifical intelligence and so on.In this paper, based on AFS theory, and the discussion of lattice theory, Boolean matrix, binary relation and their relationship, we get an algorithm to decomposite the complex comcepts into simple concepts, and apply the decomposition algorithm into diagnosis of complex faults. The algorithm in this paper is easy to understand and interpete, so it is convenience to be applied. It is expected to be widely applied in machine learning and pattern recognition.The limits of this algorithm are that we sort the sample and decomposite the complex concepts based on human intuition. But the concepts are different when the people are different or the condition is different. So we need to find some common rules to make the machine study and recognize. The combination of the simple concepts and rules make sure the machine recognize the complex concept in any condition which need to do more study and research.