基于ISODATA的电力负荷曲线分类

迭代自组织数据分析算法（ISODATA）是一种基于统计模式识别的非监督学习动态聚类算法。针对当前各算法初始聚类数取值困难、容易陷入局部最优等问题，介绍了ISODATA的原理和实现步骤，并将此算法应用于负荷分类中。在MATLAB中结合具体日负荷曲线样本进行聚类分析，结果证明聚类效果较好。将ISODATA与各种传统聚类方法进行了对比实验，比较各种算法的聚类效果、预定聚类数目对算法结果的影响，以及初始聚类中心的选择对结果的影响。对比结果证明，此方法适用于负荷分类的研究。

A Load Curve Clustering Algorithm Based on ISODATA

Interative self-organization data analysis algorithm(ISODATA) is a kind of unsupervised learning dynamic clustering algorithm based on statistical pattern recognition.The principle and process of the ISODATA are introduced in detail.This algorithm is applied in load classification,combined with specific daily load curve samples in matlab for cluster analysis.ISODATA is compared with traditional clustering methods in the results of algorithm clustering and the effects of the clustering center are better.The result is compared with the traditional clustering effect of various clustering methods and proves to be relatively more accurate,which confirms this verified the applicability of method in the study of the load curve classification.