论文部分内容阅读
不确定性是人工智能中的研究热点,而熵理论和信息粒度是度量信息系统不确定性的两种主要方法。本文在给出信息系统中熵的物理意义基础上,提出了信息系统中熵增加原理,给出了熵和信息粒度的公理化定义,据此构造了两种新的不确定性度量函数—α熵和α粒度,它们是已有熵和粒度度量的进一步推广。研究表明,α熵以及现有的熵度量都是广义熵的特殊形式,而α粒度以及现有的粒度度量都是广义粒度的特殊形式,从而统一、规范和发展了完备与不完备信息系统中的不确定性度量方法。
Uncertainty is the research hotspot in artificial intelligence, and entropy theory and information granularity are the two main methods to measure the uncertainty of information system. In this paper, based on the physical meaning of entropy in information system, the principle of entropy increase in information system is proposed, and the axiomatic definition of entropy and information granularity is given. Based on this, two new measurement functions of uncertainty Entropy and α-granularity, which are further generalizations of the existing measures of entropy and granularity. The results show that α-entropy and existing entropy measures are both special forms of generalized entropy. Both α-granularity and existing granularity measure are special forms of generalized granularity, so that the complete and incomplete information systems are unified, standardized and developed Uncertainty measurement method.