摘要
为了解决高炮系统防空效能仿真中战技指标的评价问题，本文提出了将神经网络学习知识转换为符号化特征表示示例，并通过分类决策树方法归纳为规则的知识自动获取方法，设计并实现了一个基于神经网络和符号归纳学习的高炮防空效能评价知识获取系统──EKANS，给出了分类决策树的链式数据结构和实现算法。系统兼有计算模型的精确性和符号模型的易理解性两大优势。目前该系统已研制成功。
To solve the problem of evaluation of technical data for the efficiency sim-ulation of antiaircraft gun system operation,this paper presents a method of automatic acquisition which translates the learned knowledge of neural network into examples of symbolic feature representation and acquires the concept rules through an approach of decision tree induction,thus forming an automatic acquisition system for the evaluation of efficiency of antiaircraft gun system operation,based on neural network and symbol-ic inductive learning—EKANS.It provides a link data structure and algorithm to im-plement this example decision tree.This system has the accuracy of its computational model and the understandability of a symbolic model.The system has been successfully developed.
出处
《兵工学报》
EI
CAS
CSCD
北大核心
1995年第4期61-65,共5页
Acta Armamentarii
关键词
高炮防空效能评价知识获取系统
神经网络
仿真
evaluation knowledge automatic acquisition,inductive learning,neural network,antiaircraft gun simulation
