欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对反导目标识别系统需对具有不同重要性的目标类别进行不同精度识别这一问题,提出一种基于直觉模糊核匹配追踪的目标识别方法.该方法利用核匹配追踪算法及平方间隔损失函数的优势设计了基于平方间隔损失函数的直觉模糊核匹配追踪学习机,并进一步扩展到任意非平方间隔损失函数,建立基于任意损失函数的直觉模糊核匹配追踪学习机.提出直觉模糊参数函数化的选取算法,使学习机的最终判决对指定的重要目标类别达到较高的识别精度.线性、同心圆样本高精度识别的实验分别展示了直觉模糊核匹配追踪算法的优越性,实际目标样本的识别结果验证了该方法的高识别率.
英文摘要：
In order to classify the target samples of diverse importance much more precisely according to the predefined importance of samples in antimissile system,a technique for target recognition based on intuitionistic fuzzy kernel matching pursuit（IFKMP） is proposed.With the help of the advantages of kernel matching pursuit（KMP） and least-squares loss function,intuitionistic fuzzy kernel matching pursuit learning machine（IFKMPLM） is devised and built based on least-squares loss function,which is expanded to any non-squares loss function,following with IFKMPLM based on that build.Ascribing to transforming intuitionistic fuzzy parameter into a renewed form of intuitionistic fuzzy function,the algorithm for ascertaining intuitionistic fuzzy function is also presented,serving for the appointed important target samples possessing much more precise.An extensive experimental comparison among KMP,FKMP,and IFKMP performed on linear and homocentric circle samples,respectively,shows its superiority over others,moreover,the recognition results of the practical target samples prove the high efficiency of the proposed technique.
同期刊论文项目
同项目期刊论文
期刊信息
位置：