Javascript must be enabled for the correct page display

Fraud Detection in Energy Delivery Using Adaptive Hankel Matrices in Classification

Loon, Sebastiaan Van (2018) Fraud Detection in Energy Delivery Using Adaptive Hankel Matrices in Classification. Master's Thesis / Essay, Computing Science.


Download (2MB) | Preview
[img] Text
Restricted to Registered users only

Download (95kB)


In the Netherlands there is a high correlation between theft of electricity and cannabis growing operations. Growing rooms and the related electricity theft pose a risk to the physical safety of the general population due to illegal manipulation of the electricity network, faulty connections, and excessive consumption of electricity. It is estimated that in the Netherlands yearly 200 million euro of losses are caused by electricity theft by cannabis growing opera-tions. Coteq is a Dutch Distribution Network Operator that tries to dismantle cannabis growing operations in collaboration with local law enforcement. In order to locate growing operations, a software solution provided by ValueA is used for the manual identification of patterns in electricity usage indicating illegal activities. This thesis provides an extension to the software solution of ValueA by automating the identification of suspicious patterns. In this thesis three dissimilarity measures are compared. Euclidean distance, Hankel based dissimilarity, and Dynamic Time Warping. Of this three measures the dynamic time warping is shown to give the best classification results.

Item Type: Thesis (Master's Thesis / Essay)
Supervisor name: Biehl, M. and Mohammadi, M.
Degree programme: Computing Science
Thesis type: Master's Thesis / Essay
Language: English
Date Deposited: 30 Aug 2018
Last Modified: 06 Sep 2018 14:04

Actions (login required)

View Item View Item