摘要
在美元真币信息充分的条件下,根据多元正态分布理论及方法建立了真币的性质模型.该模型通过存储正态分布的参数有效保留了美元的信息,减少存储空间,提高检索速度.提出了基于联合概率的真伪识别方法,考虑了传感器获取的所有信息,能够消除维数的影响,对指导货币性质研究及识别方法选取有实际的应用价值.
Based on the theory of multivariate normal distribution, a novel way to evaluate the effective and precise sensors used in currency recognition is proposed. The model of currency attribute is built and used to save the information of currency attribute by storing the parameter of probability function, which can decreases the memory for storing information and improve the speed of searching. It has considered all the information that the sensors gains and reduced the dimension effect. So this way has a broadly value for the study of currency property and the selection of recognition method.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2010年第9期1710-1715,共6页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(60425310)
湖南省自然科学基金(05JJ40103)
湖南省社会科学基金(09ZDB080)
关键词
货币模型
多元正态分布
货币识别
传感器
currency model
multivariate normal distribution
currency recognition
sensor
