摘要
针对电力工程数据校核准确度低的问题,提出一种结合两阶段模糊C均值聚类(TSFCM)、多种群粒子群优化支持向量机(MPSO-SVM)的电力工程小样本数据智能校核模型。利用TSFCM算法对输入的电力工程历史数据进行分类,减少不同类别电力工程特征差异对数据校核结果的影响。通过MPSO-SVM算法分别对每类电力工程数据进行校核模型构建,实现电力工程数据的精准校核。以造价数据为样本进行的仿真分析结果表明,所提出的TSFCM算法具有更优的聚类精度和计算速度,所提出的MPSO-SVM算法具有更小的校核误差,相比传统SVM算法,平均校核误差由9.3%下降到4.7%。
To solve the problem of low accuracy of power engineering data verification,this paper proposes an intelligent verification model for small sample data of power engineering.The model combines two-stage fuzzy C-means clustering(TSFCM)with multi-population particle swarm optimization support vector machine(MPSO-SVM).This model uses the TSFCM algorithm to classify the input historical data of power engineering to reduce the impact of differences in the characteristics of different types of power engineering features on data verification results.The MPSO-SVM algorithm is used to construct a verification model for each type of power engineering data to achieve accurate verification of power engineering data.The simulation analysis results using cost data as samples show that the proposed TSFCM algorithm has better clustering accuracy and computational speed,and the proposed MPSO-SVM algorithm has smaller verification error.Compared with traditional SVM algorithm,the average verification error decreases from 9.3%to 4.7%.
作者
田海丰
华生萍
杨蒲寒婷
才海多杰
刘舒宁
TIAN Haifeng;HUA Shengping;YANGPU Hanting;CAIHAI Duojie;LIU Shuning(Economic and Technological Research Institute of State Grid Qinghai Electric Power Company,Xining 810008,China)
出处
《微型电脑应用》
2025年第3期82-85,89,共5页
Microcomputer Applications
基金
国网青海省电力公司管理研究类项目(Q2021RCDT2B0713)。
关键词
电力工程数据
多种群粒子群
支持向量机
数据校核
C均值聚类
power engineering data
multi-population particle swarm
support vector machine
data verification
C-means clustering
