Improved optimization feature-based unsupervised abnormal power consumption behavior detection method

• Main Authors: QIN YUYI, LIAO KAIJI, CAI JUNYI, LI QILIN
• Format: Patent
• Language: Chinese; English
• Published: 05.07.2024
• Subjects: CALCULATING; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention discloses an unsupervised abnormal power consumption behavior detection method based on improved optimization characteristics. The method comprises the following steps: data acquisition and preprocessing; constructing and selecting a feature data set; unsupervised feature selection based on improved maximum correlation and minimum redundancy; using density-based noise to apply spatial clustering anomaly detection; and abnormal category judgment based on dynamic time warping. The detection method provided by the invention has higher universality, ensures normal operation of an algorithm by modifying initial parameters for different regions, extracts features with higher value in user data based on variance and improved unsupervised feature selection, eliminates data with little help to power consumption behavior judgment, and improves the accuracy of power consumption behavior judgment. Compared with the prior art, the method has the advantages that high-value feature data are concentrated, the f