多传感器的BPNN和SVM多源异构数据融合算法
Multi Sensor Heterogeneous Data Fusion Algorithm Based on BPNN and SVM
王晓琪 1陈颖聪 1谢敏敏 1张嘉慧 1蔡上1
作者信息
- 1. 广东电网有限责任公司 梅州供电局,广东 梅州 514021
- 折叠
摘要
多传感器的多源异构数据融合处理时,大量的冗余数据及复杂的非线性可分空间导致能耗较大,为此,提出了BP神经网络和支持向量机的多源异构数据融合算法.以数据关系构建约束条件,利用 BP神经网络算法建立数据清洗模型,判定节点变量的活跃程度,优化数据输入;建立数据集合,提取数据特征向量;利用支持向量机泛化能力强、凸优化的特点,获取特征的最优分类超平面,获得非线性可分多源数据集转化为高维线性可分空间的最优决策值,输出结果.实验结果表明,该算法融合多源异构数据的能量消耗小、延迟低,融合效果好.
Abstract
In the process of multi-sensor multi-source heterogeneous data fusion processing,a large number of redundant data and complex nonlinear separable space lead to high energy consumption.Therefore,a multi-source heterogeneous data fusion algorithm based on BP neural network and support vector machine is proposed.Based on the data relationship,the constraint conditions were established,and the BP neural network algorithm was used to establish the data cleaning model,and the activity degree of node variables was determined to optimize the data input.To set up data set and extract data fea-ture vector;Based on the support vector machine's strong generalization ability and convex optimization,the optimal classi-fication hyperplane of the features is obtained,and the optimal decision value of the nonlinear separable multi-source data set is obtained into the high-dimensional linear separable space.Experimental results show that this algorithm has low energy consumption,low delay and good fusion effect.
关键词
BP神经网络/支持向量机/多源异构数据/数据清洗/数据融合Key words
BP neural network/support vector machine/multi source heterogeneous data/data cleaning/data fusion引用本文复制引用
基金项目
国家电网新一代人工智能科技项目(2020AAA0103400)
出版年
2024
国家科技期刊平台
NSTL
万方数据