Scholarlink

The purpose of this study is to analyze electricity consumption data from 100 customers in order to determine the optimal clustering technique for smart electricity meter data. Our three-cluster validity indices are the silhouette index, the Dunn index, and the covariance-based index which allows the identification of multiple optimal clusters. One of three methods—K-means hierarchical clustering deep clustering and others—is then used to cluster the dataset. To determine whether these methods are appropriate we examine the clustering results using the silhouette index and Within-Cluster Sum of Squares (WCSS). In order to provide insights into the best strategy for data-driven energy management, our study suggests the best clustering technique for dividing up smart electricity meter data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advanced Computing Techniques in Electrical Engineering Assessing Machine Learning Approaches for Load Pattern Analysis in Low-Voltage Distribution System

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • M. Muthamizh Selvam,
  • K. Kishore Babu,
  • P. G. Seetharam,
  • V. Kavitha,
  • G. Rajalakshmi,
  • Jaishree

摘要

The purpose of this study is to analyze electricity consumption data from 100 customers in order to determine the optimal clustering technique for smart electricity meter data. Our three-cluster validity indices are the silhouette index, the Dunn index, and the covariance-based index which allows the identification of multiple optimal clusters. One of three methods—K-means hierarchical clustering deep clustering and others—is then used to cluster the dataset. To determine whether these methods are appropriate we examine the clustering results using the silhouette index and Within-Cluster Sum of Squares (WCSS). In order to provide insights into the best strategy for data-driven energy management, our study suggests the best clustering technique for dividing up smart electricity meter data.