科技通报
科技通報
과기통보
BULLETIN OF SCIENCE AND TECHNOLOGY
2015年
2期
146-148
,共3页
蚁群算法%特征优选%信息融合
蟻群算法%特徵優選%信息融閤
의군산법%특정우선%신식융합
ant colony algorithm%feature selection%information fusion
在网络安全预测监护模型设计中，需要对网络安全监护信息进行数据融合和特征优选，以提高对变异特征的识别能力。传统方法中，采用蚁群算法进行监护信息特征优化融合进化和链路模型设计，算法无法实现相邻簇头之间的信息素融合，特征优化效果不好。针对这一问题，提出蚁群链运动多层博弈的网络监护信息融合特征优选算法，构建多层博弈网络监护数据样本驱动空间权矩阵模型，引入粗糙集理论，对蚁群引导的粗糙集前馈补偿网络进行动态博弈，实现网络安全监护数据的预测控制目标函数最佳寻优。构建多层博弈网络监护系统模型，得到蚁群链运动的监护信息数据状态跟踪模型，实现网络安全监护信息的融合特征优选改进。仿真实验表明，该算法能有效提高对异常信息的监护和检测能力，有较高的特征优选品质，展示了本文算法在对网络安全监护中的优越性能。
在網絡安全預測鑑護模型設計中，需要對網絡安全鑑護信息進行數據融閤和特徵優選，以提高對變異特徵的識彆能力。傳統方法中，採用蟻群算法進行鑑護信息特徵優化融閤進化和鏈路模型設計，算法無法實現相鄰簇頭之間的信息素融閤，特徵優化效果不好。針對這一問題，提齣蟻群鏈運動多層博弈的網絡鑑護信息融閤特徵優選算法，構建多層博弈網絡鑑護數據樣本驅動空間權矩陣模型，引入粗糙集理論，對蟻群引導的粗糙集前饋補償網絡進行動態博弈，實現網絡安全鑑護數據的預測控製目標函數最佳尋優。構建多層博弈網絡鑑護繫統模型，得到蟻群鏈運動的鑑護信息數據狀態跟蹤模型，實現網絡安全鑑護信息的融閤特徵優選改進。倣真實驗錶明，該算法能有效提高對異常信息的鑑護和檢測能力，有較高的特徵優選品質，展示瞭本文算法在對網絡安全鑑護中的優越性能。
재망락안전예측감호모형설계중，수요대망락안전감호신식진행수거융합화특정우선，이제고대변이특정적식별능력。전통방법중，채용의군산법진행감호신식특정우화융합진화화련로모형설계，산법무법실현상린족두지간적신식소융합，특정우화효과불호。침대저일문제，제출의군련운동다층박혁적망락감호신식융합특정우선산법，구건다층박혁망락감호수거양본구동공간권구진모형，인입조조집이론，대의군인도적조조집전궤보상망락진행동태박혁，실현망락안전감호수거적예측공제목표함수최가심우。구건다층박혁망락감호계통모형，득도의군련운동적감호신식수거상태근종모형，실현망락안전감호신식적융합특정우선개진。방진실험표명，해산법능유효제고대이상신식적감호화검측능력，유교고적특정우선품질，전시료본문산법재대망락안전감호중적우월성능。
In the network security forecast monitoring model design, it needs for data fusion and feature selection of network security monitoring information, in order to improve the recognition ability of variability. In the traditional method, the ant colony algorithm for monitoring information optimization fusion evolution and link model design, the algorithm cannot be achieved between the adjacent cluster heads of pheromone fusion, and feature optimization effect is not good. Aiming at this problem, the network monitoring information fusion and feature selection algorithm is proposed based on ant colony chain motion multi game, a multilayer monitoring data of game network sample driven spatial weight matrix model is constructed, the rough set theory is introduced to guide the ant colony, rough set feed forward compensation network is a dynamic game, it is used to forecast the network security monitoring data and control the objective function optimization. Multi game net?work monitoring system model is constructed, the information monitoring data state ant chain motion tracking model is ob?tained, and the integration of feature selection of network security monitoring information is completed. Simulation results show that, the algorithm can effectively improve the abnormal information monitoring and detection capability, feature selec?tion of higher quality is improved, it shows the algorithm is superior in performance of the network security monitoring.
