基于属性加权贝叶斯分类器的目标毁伤效果分析

Battle Damage Assessment Based on Attribute Weighted Bayesian Classification

目标毁伤效果评估(BDA)是作战决策的重要组成部分,信息化条件下的现代战争给目标毁伤评估增加了复杂性与不确定性。贝叶斯网络是最近几十年流行起来的一种不确定性推理工具,朴素贝叶斯分类是一种简单而高效的方法,但是它的属性独立性假设,影响了它的分类性能。通过对朴素贝叶斯分类器进行属性加权的改进,可以提高分类效果,并举例演示了利用改进贝叶斯分类器进行目标BDA的过程,说明了这种改进的朴素贝叶斯分类在目标BDA中应用的可行性与有效性。

Target battle damage assessment（BDA）is an important part of combat decision.Information-based modern war increases the complexity and uncertainty of BDA.Though Bayesian network is a popular way to uncertainty reason in recent decades and Nave Bayesian Classification is simple and efficient,its hypothesis of attribute indepentence impacts the classification performance.Through attribute weighted Nave Bayesian Classification,the performance of Nave Bayes has been improved.And with an example of BDA procedure,the improved Bayesian Classification shows its feasibility and effectiveness in target BDA.