Title:
NEW NON-PARAMETRIC STATISTICAL BEHAVIOR IDENTIFICATION ECOSYSTEM FOR ELECTRIC-POWER ILLEGAL USE DETECTION
Document Type and Number:
Japanese Patent JP2023029983
Kind Code:
A
Abstract:
To provide a new non-parametric statistical behavior identification ecosystem and a method for electric-power illegal use detection.SOLUTION: An electric-power illegal use detection system includes a behavior detection ecosystem to improve a detection rate of electric-power illegal use while reducing an error determination rate, in which machine learning algorithms are eschewed in favor of two separate models that are applied sequentially. A first model improves the detection rate of electric-power illegal use by using a detector to identify a customer engaging in questionable behavior on the basis of a demand profile of each customer. A second model reduces the error determination rate by identifying potentially legitimate explanation for any questionable behavior. Subtracting suspicious behavior with a legitimate explanation leaves only identified but unexplained questionable behavior that is most likely associated with an illegal act.SELECTED DRAWING: Figure 2A
Inventors:
Abbas, Hussein
Application Number:
JP2022195769A
Publication Date:
March 07, 2023
Filing Date:
December 07, 2022
Export Citation:
Assignee:
Oracle International Corporation
International Classes:
G06Q50/06
Attorney, Agent or Firm:
Patent Attorney Fukami Patent Office