Masters Thesis Defense for Mohammad Khan • 06/13/18 at 1:00 PM
May 25, 2018
Dear Faculty, Graduate and Undergraduate students,
You are cordially invited to my thesis defense.
Title: Non parametric learning for energy disaggregation.
When: Wednesday, June 13, 2018, at 1:00 pm.
Where: Simrall Hall, Room 228 (Conference Room)
Candidate: Mohammad Mahmudur Rahman Khan
Degree: Masters of Science, Electrical and Computer Engineering
Committee:
Dr. Bo Tang
(Major Professor)
Dr. John E. Ball
(Committee Member)
Dr. Bryan A. Jones
(Committee Member)
Abstract:
It is possible to reduce the energy consumption if consumers get an item-wise, appliance-by-appliance electricity bill rather than the bill which provides information about aggregated energy consumption. Though some modern appliances are capable of communicating with utilities, older appliances do not have that ability. Therefore, for monitoring appliance-wise energy consumption we need to disaggregate the total electricity load into individual appliance loads.
This thesis work presents a non-parametric learning method, Extended Nearest Neighbor (ENN) algorithm, as a tool for data disaggregation in smart grid. The ENN algorithm makes prediction according to the maximum gain of intra-class coherence. This algorithm not only considers the K nearest neighbors of the test sample but also considers whether these K data points consider the test sample as their nearest neighbor or not, whereas traditional K-Nearest Neighbor (KNN) algorithm only relies on the classes of the K nearest neighbors of the test sample. So far, ENN has shown noticeable improvement in the classification accuracy for various real-life applications. To further enhance its prediction capability, in this thesis we propose to incorporate a metric learning algorithm, namely the Large Margin Nearest Neighbor (LMNN) algorithm, as a training stage in ENN. Our experiments on real-life energy data sets have shown significant performance improvement compared to several other traditional classification algorithms, including KNN and Support Vector Machines.
Regards,
Mohammad Mahmudur Rahman Khan
Graduate Teaching Assistant
Department of Electrical and Computer Engineering
Bagley College of Engineering
Mississippi State University