Method and system for cleaning and classifying abnormal power utilization data of power consumers

The invention discloses a method and a system for cleaning and classifying abnormal power consumption data of power consumers. The method comprises the following steps: collecting historical power consumption data of a power consumer; preprocessing the historical electricity consumption data; expanding the preprocessed historical electricity consumption data to obtain a data set; constructing an electricity larceny discriminator; and classifying the electricity consumption data of the monitored power users in the power distribution network in real time by using the trained electricity stealing discriminator, and judging whether the monitored power users are non-electricity stealing users or electricity stealing users according to a classification result. According to the method, a scheme of expanding the historical electricity consumption data set is adopted, minority-class synthetic samples are added, redundant majority-class samples are deleted, balanced distribution of the data set is realized, the problem