Adaptive Momentum Neural Learning for Anomaly Detection of Power Distribution and Consumption Load Data

DOI

10.2991/meici-18.2018.212How to use a DOI?

Keywords

Anomaly detection; Adaptive momentum; Neural network; Power distribution and consumption

Abstract

In order to overcome the high misstatement rate and false positive rate brought by the traditional anomaly detection technologies on power distribution and consumption load data, in this paper, a novel momentum adaptive neural learning-based anomaly detection method for power distribution and consumption load data is proposed by adding momentum adaptive mechanism into the error back propagation (BP) model. In the proposed algorithm, according to the gradient changes of experience generalization error in the anomaly detection process, the momentum factor is adjusted in the BP algorithm to avoid the training results of neural network learning converges to a local minimum value. The contrastive simulation experiments on real-time database of power distribution and consumption load data demonstrate that, our proposed method holds better performances on accuracy, misstatement rate and false positive rate, compared with other traditional algorithms.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Yan Liu
PY  - 2018/12
DA  - 2018/12
TI  - Adaptive Momentum Neural Learning for Anomaly Detection of Power Distribution and Consumption Load Data
BT  - Proceedings of the 2018 8th International Conference on Management, Education and Information (MEICI 2018)
PB  - Atlantis Press
SP  - 1059
EP  - 1064
SN  - 1951-6851
UR  - https://doi.org/10.2991/meici-18.2018.212
DO  - 10.2991/meici-18.2018.212
ID  - Liu2018/12
ER  -