一个改进的N元自适应模式识别系统开发工具

An Improved Development Tool for N-tuple Adaptive Pattern Recognition Systems

N元自适应模式识别系统开发工具NIRSS是一个通用开发工具.它可以对多种N元自适应模式识别系统模型的体系结构和应用对象进行试验研究.本文针对该系统的不足,提出采用修改编译模式并利用EMS技术扩大存贮容量的方法,使其具有处理高维数模式数据(如图像)的能力.实验证明,改进工作是成功且有意义的.

NIRSS is a general-purpose development tool for N-tuple adaptive pattern recognition systems. It can be used to research the architecture and application of many models of TV-tuple adaptive pattern recognition system . In order to overcome the disadvantages of this system, we propose the methods of modifing the compile model and expanding the memory capacity with the EMS techniguc to improve the performance of the original NIRSS. In this way, the improved system may process the large dimensional pattern data (e. g. image pattern). The experiments show that the improving task is successful and significant.