摘要
针对支持向量聚类算法训练样本不稳定问题,引入数据场概念,提出一种基于数据场的支持向量聚类算法,将数域空间构成的数据场中势值较高样本作为训练集获得模型再进行预测聚类。将改进的算法用于态势估计中目标分群问题,仿真结果表明:该算法在样本容量不是很高条件下的准确率较传统算法有所提高,很好地解决了训练样本选择影响聚类效果的问题,但改进后算法耗时较原先有所增加。
Considering the Support Vector Clustering（SVC）algorithm to the problem of unstable training sample,with the concept of data field,an improved algorithm based on data field is proposed（Data Field Support Vector Clustering,DFSVC）,the data sample with higher potential value from the data field in composition of the sample space is regarded as the training sample for training the model,and then predict the cluster with the model. The improved algorithm is applied to target grouping of situation assessment. The simulation shows that this algorithm accuracy is higher than the traditional algorithm under the condition of samle size is not very large,it is good to solve the problem of the influence of clustering results with the training sample selection. The time-consuming of improved algorithm is higher than original algorithm.
出处
《火力与指挥控制》
CSCD
北大核心
2015年第12期40-43,共4页
Fire Control & Command Control
基金
国家自然科学基金面上项目(61179036)
关键词
目标分群
支持向量聚类
数据场
势函数
target grouping
support vector clustering
data field
potential function
