An electric price inspection execution method based on deep learning of big data

Abstract

The invention belongs to the technical field of electric power marketing inspection, in particular to an electric price inspection execution method based on deep learning of big data, which is an innovative electric price execution inspection method in which the variational automatic encoder algorithm in the deep learning is used for identifying abnormal electric price execution. The invention comprises the following steps: acquiring electric power data and simple classification; extracting probabilistic features from power consumption data; reconstructing probability discrimination. The invention effectively solves the calculation problem of a plurality of heterogeneous parameters in the current electricity price execution inspection, and effectively judges whether the power consumption customer is abnormal by reconstructing probability. Aiming at the partial missing of power consumption data of some customers, the problem of missing data can be solved effectively because the model generated in the process of algorithm discrimination can recover the same feature of the data. The on-line diagnosis of electricity price execution inspection is realized, which solves the practical problems of diversified electricity price execution and complex parameters, and provides effective guarantee for electric power marketing work, greatly improves the accuracy of abnormal detection, and greatly reduces the mismatch rate of inspection.

Description

technical field [0001] The invention belongs to the technical field of electric power marketing inspection, and in particular relates to an electricity price inspection execution method based on big data deep learning, which is an innovative method for electricity price inspection by using a variational autoencoder algorithm in deep learning to identify abnormalities in electricity price execution . Background technique [0002] The marketing inspection of power supply enterprises is an important guarantee for the smooth development of marketing business, timely discovery of abnormal situations, and improvement of work quality and management level. In recent years, the scale of smart grids has become larger and larger, and big data, multi-dimensionality, high intelligence, and strong reliability have become the salient features of modern power grids. Traditional inspection and monitoring work methods are difficult to adapt to the needs of new forms, and intelligent methods ...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problems existing in the above-mentioned prior art, the present invention provides an electricity price inspection execution method based on big data deep learning. It is difficult to comprehensively inspect the implementation of electricity prices for electricity consumption

Images