Scholarlink

<p>This paper presents a two-stage network structure for evaluating electricity distribution companies (EDCs) and defines appropriate variables for each stage. A suitable network data envelopment analysis (NDEA) model is then applied to measure the overall and sub-system efficiencies. The sub-systems are defined as the “production stage” and the “profitability stage.” To reduce data dimensionality and extract independent variables, principal component analysis (PCA), a popular machine learning technique, is applied. Because some principal components are discarded, information loss occurs, which introduces uncertainty into the NDEA model. To model this type of uncertainty, fuzzy theory is used, and the proposed NDEA model is developed in a fuzzy environment. The output of the PCA model is treated as the uncertain data for the NDEA model. To illustrate the capability of the proposed PCA-NDEA approach, a case study consisting of 39 electricity distribution companies (EDCs) is investigated, and the overall efficiency as well as the efficiencies of Stages 1 and 2 are calculated.</p>

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

Modeling PCA information loss with fuzzy theory in a network DEA framework: efficiency evaluation of electricity distribution companies

  • 期刊论文

  • 发表日期：

    2025年12月2日

查看全文
  • Hashem Omrani,
  • Arezoo Sheikhani,
  • Raha Imanirad

摘要

This paper presents a two-stage network structure for evaluating electricity distribution companies (EDCs) and defines appropriate variables for each stage. A suitable network data envelopment analysis (NDEA) model is then applied to measure the overall and sub-system efficiencies. The sub-systems are defined as the “production stage” and the “profitability stage.” To reduce data dimensionality and extract independent variables, principal component analysis (PCA), a popular machine learning technique, is applied. Because some principal components are discarded, information loss occurs, which introduces uncertainty into the NDEA model. To model this type of uncertainty, fuzzy theory is used, and the proposed NDEA model is developed in a fuzzy environment. The output of the PCA model is treated as the uncertain data for the NDEA model. To illustrate the capability of the proposed PCA-NDEA approach, a case study consisting of 39 electricity distribution companies (EDCs) is investigated, and the overall efficiency as well as the efficiencies of Stages 1 and 2 are calculated.