Principal Component and Physical Quantities Enhanced Logistic Regression for Electric Load Classification

The classification and recognition of household appliance loads is of great significance for understanding user electricity consumption behavior, optimizing electricity planning, and intelligent management of the power grid, providing data support for efficient operation of the power grid and implementation of energy policies. This paper proposes a principal component and physical quantities enhanced logistic regression for electric load classification method. The data processing method using principal component analysis and physical quantities extraction can effectively reduce dimensionality while preserving the main information in the data, improving the performance and interpretability of the model, reducing computational and storage costs, and is suitable for the analysis and processing of high-dimensional data. Case study shows that using this method for load classification with only 3.98% of the original data volume, the overall classification accuracy can reach <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 1. 8 4 \%}$</tex>.