摘要
传统舰船通信系统非平衡数据检测技术需要计算所有数据,检测效率低下。对此设计舰船通信系统非平衡数据多目标自动检测技术。通过计算通信数据集内各数据点的欧几里德距离,剔除带有明显维度特征的非平衡数据减少计算量,将剩余通信数据降维并建立一维空间映射,最后对空间映射内的一维数据集特异值进行统计计算,实现非平衡数据检测。实验研究表明相比较传统检测方式,单位时间下设计的多维数据异常点检测数量高25%,固定异常点数据用时缩减23%,可以明显提高检测效率。
The non-equilibrium data detection technology of traditional shipboard communication system needs to calculate all the data and the detection efficiency is low. The multi-target automatic detection technology for non-balanced data of the ship communication system is designed. By calculating the Euclidean distance of each data point in the communication data set, the calculation amount of non-equilibrium data with obvious dimensional characteristics is eliminated, the dimension of remaining communication data is reduced and one-dimensional space map is established. At last, the specific value of one-dimensional data set in the spatial map is statistically calculated to realize non-balanced data detection. Experimental studies have shown that compared with traditional detection methods, the number of multi-dimensional data abnormal point detection designed under unit time is 25% higher, and the time of fixed abnormal point data is reduced by 23%,which can significantly improve the detection efficiency.
作者
王玉华
WANG Yu-hua(Hulunbuir Vocational Technical College, Hulunbuir 021000, China)
出处
《舰船科学技术》
北大核心
2018年第12X期127-129,共3页
Ship Science and Technology
关键词
通信系统
平衡数据
自动检测
communication systems
balanced data
automatic detection