Electricity Theft Detection using Fusion DenseNet-RF Model
AUTHORS
Tai-hoon Kim, University of Tasmania， Churchill Ave, Hobart TAS 7005, Australia
ABSTRACT
Aiming at the problems of high cost and low efficiency in power theft detection, this paper proposes a new type of DenseNet-RF model to detect power users theft. The DenseNet and Random Forest (RF) algorithms are fused, where DenseNet is used to automatically extract customer power usage characteristics, and RF is used to classify customer power abnormalities. By introducing the SMOTE algorithm, the imbalance problem of customer power consumption data was solved, the hyperparameters inside DenseNet were adjusted and the network model was pruned to obtain a better preprocessing model. When training the random forest classifier, call the preprocessing model and optimize the classifier parameters through grid search, and then obtain the final fusion algorithm model. The experimental results show that the DenseNet-RF fusion model effectively improves the accuracy of classification. Compared with the ensemble learning algorithm in single machine learning, traditional convolutional neural network, DenseNet, and algorithms of existing research results, the algorithm adopted by this model has better classification accuracy and stability, and the model has good generalization. ability.
KEYWORDS
Random forest, DenseNet, Electricity stealing detection, Network search algorithm
REFERENCES
[1] S. Mclaughlin, D. Podkuiko, and P. McDaniel, “Energy theft in the advanced metering infrastructure,” Proceedings of the 2009 International Workshop on Critical Information Infrastructures Security. Cham: Springer, pp.176-187, (2009)
[2] H. K. Lee and K. Nam, “On-line overshoot suppression method for EV propulsion motor considering cross-coupled inductance,” IEEE Transactions on Industrial Electronics, pp.9255-9265, (2018)
[3] A. H. Nizar, Z. Y. Dong, and Y. Wang, “Power utility non-technical loss analysis with extreme learning machine method,” IEEE Transactions on Power Systems, vol.23, no.3, pp.946-955, (2008)
[4] J. Nagi, K. S. Yap., and S. K. Tong, “Nontechnical loss detection for metered customers in power utility using support vector machines,” IEEE Transactions on Power Delivery, vol.25, no.2, pp.1162-1171, (2009)
[5] R. R. Bhat, R. D. Trevizan, and R. Sengupta, “Identifying non-technical power loss via spatial and temporal deep learning,” Proceedings of the 2016 IEEE International Conference on Machine Learning and Applications. Piscataway: IEEE, pp.272-279, (2016)
[6] P. McDaniel and S. McLaughlin, “Security and privacy challenges in the smart grid,” IEEE Security and Privacy, vol.7, no.3, pp.75-77, (2009)
[7] G. Pleiss, D. Chen, and G. Huang, “Memory-efficient implementation of DenseNets,” Computer Vision and Pattern Recognition, vol.2, no.1, pp.47-49, (2017)
[8] D. Anguita, A. Ghio, and N. Greco, “Model selection for support vector machines: Advantages and disadvantages of the machine learning theory,” Neural Networks (IJCNN), The 2010 International Joint Conference on. IEEE, (2010)
[9] Z. Khandezamin, M. Naderan, and M. J. Rashti, “Detection and classification of breast cancer using logistic regression feature selection and GMDH classifier,” Journal of Biomedical Informatics, (2020)
[10] M. Balachandran, T. H. Shin, and M. O. Kim, “AIPpred: Sequence-based prediction of anti-inflammatory peptides using random forest,” Frontiers in Pharmacology, vol.9, pp.276, (2018)
[11] M. Oprescu, V. Syrgkanis, and Z. S. Wu, “Orthogonal random forest for heterogeneous treatment effect estimation,” Vasilis Syrgkanis, (2018)
[12] F. Iandola, M. Moskewicz, and S. Karayev, “DenseNet: Implementing efficient ConvNet descriptor pyramids,” Eprint Arxiv, (2014)
[13] B. Fielding and Z. Li, “Evolving deep DenseBlock architecture ensembles for image classification,” Electronics, vol.9, no.11, pp.1880, (2020)
[14] L. Breiman, “Random forests,” Machine Learning, vol.45, no.1, pp.5-32, (2001)
[15] L. Breiman, “Bagging predictors,” Machine Learning, vol.24, no.2, pp.123-140, (1996)
[16] D. Kushary, “Bootstrap methods and their application,” Technometrics, vol.42, no.2, pp.216-217, (2000)
[17] P. Jokar, N. Arianpoo, and V. C. M. Leung, “Electricity theft detection in AMI using customers’ consumption patterns,” IEEE Transactions on Smart Grid, vol.7, no.1, pp.216-226, (2015)
[18] N. V. Chawla, K. W. Bowyer, and L. O. Hall, “SMOTE: Synthetic minority over-sampling technique,” Journal of Artificial Intelligence Research, vol.16, no.1, pp.321-357, (2002)