摘要
雷达辐射源的分选识别算法通常是以正确分选率或识别率为重要指标,但该指标通常是一个点估计值,无法反映系统识别的稳定性和估计精度。本文利用小子样Bootstrap法在有限的实验样本的基础上进行区间估计,并提出了一个特征分选识别率的稳定性判定因子,通过该因子能够量化系统识别率的稳定性。
The right recognition rates are the most important index of recognition or classification of radar emitter,but this point estimation can not indicate the stability of classification and the precision of estimation.This article forms the estimation at confidence interval with the Bootstrap method and presents a stability factor with which the stability of classification can be clearly indicated and quantified.
出处
《成都电子机械高等专科学校学报》
2011年第2期9-12,共4页
Journal of Chengdu Electromechanical College