成都工业学院学报
成都工業學院學報
성도공업학원학보
Journal of Chengdu Technological University
2014年
4期
11-14
,共4页
膜计算%特征选择%全局优化
膜計算%特徵選擇%全跼優化
막계산%특정선택%전국우화
membrane computing%feature selection%global optimization
提出一种基于膜计算的C波段异常信号特征选择算法。该算法将膜结构分为2层:基本膜和表层膜。首先在基本膜中结合模糊C均值( FCM)聚类、粗糙集和遗传算法( GA)实现局部寻优,然后将局部最优解传送给表层膜,最后在表层膜区域中再一次使用遗传算法实现全局优化。通过对无线电监测C波段异常信号特征库进行仿真实验,发现该算法在C波段异常信号的特征选择上比使用遗传算法、约翰逊算法所得的条件属性更简单且优化效果更好。
提齣一種基于膜計算的C波段異常信號特徵選擇算法。該算法將膜結構分為2層:基本膜和錶層膜。首先在基本膜中結閤模糊C均值( FCM)聚類、粗糙集和遺傳算法( GA)實現跼部尋優,然後將跼部最優解傳送給錶層膜,最後在錶層膜區域中再一次使用遺傳算法實現全跼優化。通過對無線電鑑測C波段異常信號特徵庫進行倣真實驗,髮現該算法在C波段異常信號的特徵選擇上比使用遺傳算法、約翰遜算法所得的條件屬性更簡單且優化效果更好。
제출일충기우막계산적C파단이상신호특정선택산법。해산법장막결구분위2층:기본막화표층막。수선재기본막중결합모호C균치( FCM)취류、조조집화유전산법( GA)실현국부심우,연후장국부최우해전송급표층막,최후재표층막구역중재일차사용유전산법실현전국우화。통과대무선전감측C파단이상신호특정고진행방진실험,발현해산법재C파단이상신호적특정선택상비사용유전산법、약한손산법소득적조건속성경간단차우화효과경호。
A feature selection algorithm based on membrane computing for C band abnormal signal is proposed in this paper.The membrane structure is divided into two layer———elementary membrane and skin membrane.Firstly, this algorithm combines with fuzzy C means clustering, rough sets and genetic algorithm to realize local optimization in elementary membrane, and then transmits the local optimal solutions to skin membrane.Finally , the genetic algorithm is also used to realize the global optimization in the skin membrane area.Through C band abnormal signal feature library of radio monitoring on the simulation experiment, results show that this algorithm can obtain simpler and better optimization results than using genetic algorithm and Johnson’s algorithm on C band abnormal signal feature selection.