Concealed electricity stealing behavior identification method based on synthetic minority class oversampling technology

Abstract

The invention discloses a hidden electricity stealing behavior identification method based on a synthetic minority class oversampling technology, and belongs to the field of power supply management. Collecting power consumption data of normal users and recorded abnormal users in a certain period of time; collecting weather temperature data of a corresponding time period; carrying out data expansion based on a minority class oversampling technology; carrying out user feature data label construction; carrying out random forest algorithm electricity stealing behavior prediction; and judging or predicting whether a user finally has an electricity stealing behavior. According to the method, except for user power utilization characteristic indexes, weather factors and user power utilization time-space correlation are considered, the hidden characteristic of power stealing behaviors is considered, abnormal samples with small sample sizes are expanded by using a synthetic minority class oversampling technology, the effective data size and coverage range of the samples are improved, the model training precision is improved, and the model training efficiency is improved. And the identification capability of the electricity stealing behavior is enhanced.

Description

technical field [0001] The invention belongs to the field of power supply management, and in particular relates to a method for identifying electricity stealing behavior by using a method based on synthetic minority class oversampling. Background technique [0002] With the rapid development of the economy, people's dependence on electricity is getting higher and higher, and the power and electricity consumption of household appliances have developed rapidly. [0003] However, some users steal electricity by changing the metering line and structure of the electricity meter in order to pay less electricity bills, or even not paying electricity bills, and electricity theft is becoming more and more rampant. [0004] Electricity theft will not only reduce the income of the power grid company, but also affect the safe and reliable power consumption of other users. Changing the line may also cause a short circuit, cause serious disasters such as fire, and threaten the safety of o...

AI Technical Summary

Problems solved by technology

[0010] The above detection and identification methods for electricity theft behavior are completely based on the user's own power data, and do not take into account additional factors such as ambient temperature, and individual users are affected by the ambient temperature, and there will be large deviations in electricity consumption behavior

[0011] In addition, the above methods are all based on the premise that there are enough samples of user electricity stealing data. In fact, electricity stealing behavior is relatively hidden, and the data of electricity stealing users is often relatively small, and sometimes it is difficult to form sufficient data. samples to evaluate

Images