摘要
随着智能电网的不断优化扩展及数据集的沉淀,海量大数据因为数据量太大、维数太高而陷入了"维数灾难"中,在工程实践中难以对其进行有效的研究。提出了运用拉普拉斯特征映射(Laplacian Eigenmaps)对电网大数据进行自适应学习并降维,运用降维后的数据在Hadoop平台上进行实验分析,证明其能有效地应用于智能电网大数据的降维运算。
With the continuous optimization and expansion of smart grid and the precipitation of data sets, the massive data is stepping into the "Curse of Dimensionality" because of the too large amount and too high dimension, which make it difficult to study the data effectively in engineering practice. This paper presents a method which employs Laplacian Eigenmaps to adaptively learn the grid big data and reduce its dimensionality, and then uses the data after dimensionality reduction for further analysis. The results of experimental analysis on the Hadoop platform show-that the proposed method can be effectively applied to dimensionality reduction of the smart grid big data.
作者
黄纯德
陈晓亮
朱珊珊
王晶华
郭光
HUANG Chunde;CHEN Xiaoliang;ZHU Shanshan;WANG JingHua;GUO Guang(Shanxi Electric Power Research Institute,Electric Power Company of State Grid,TaiYuan Shanxi 030001,China;Shanxi Electric Power Company of State Grid,TaiYuan Shanxi 030001,China;Beijing Zhongke Chuangyi Technology Co.,Ltd,BeiJing 100198,China)
出处
《计算机与网络》
2018年第18期69-71,共3页
Computer & Network
关键词
智能电网大数据
机器学习
拉普拉斯特征映射
数据降维
big data of smart grid
machine learning
Laplacian Eigenmaps
dimensionality reduction