{"references": ["C.-S. Chen, T.-T. Ku, and C.-H. Lin, \"Design of phase identification\nsystem to support three-phase loading balance of distribution feeders,\"\nIEEE Transactions on Industry Applications, vol. 48, no. 1, pp. 191\u2013198,\n2012.", "K. J. Caird, \"Meter phase identification,\" Mar. 27 2012, uS Patent\n8,143,879.", "M. H. Wen, R. Arghandeh, A. von Meier, K. Poolla, and V. O. Li, \"Phase\nidentification in distribution networks with micro-synchrophasors,\" in\n2015 IEEE Power & Energy Society General Meeting. IEEE, 2015,\npp. 1\u20135.", "M. Dilek, R. P. Broadwater, and R. Sequin, \"Phase prediction in\ndistribution systems,\" in Power Engineering Society Winter Meeting,\n2002. IEEE, vol. 2, 2002, pp. 985\u2013990.", "V. Arya, D. Seetharam, S. Kalyanaraman, K. Dontas, C. Pavlovski,\nS. Hoy, and J. R. Kalagnanam, \"Phase identification in smart\ngrids,\" in Smart Grid Communications (SmartGridComm), 2011 IEEE\nInternational Conference on, Oct 2011, pp. 25\u201330.", "H. Pezeshki and P. J. Wolfs, \"Consumer phase identification in a three\nphase unbalanced LV distribution network,\" in 2012 3rd IEEE PES\nInnovative Smart Grid Technologies Europe (ISGT Europe), Oct 2012,\npp. 1\u20137.", "T. A. Short, \"Advanced metering for phase identification, transformer\nidentification, and secondary modeling,\" IEEE Transactions on Smart\nGrid, vol. 4, no. 2, pp. 651\u2013658, June 2013.", "W. Wang, N. Yu, B. Foggo, J. Davis, and J. Li, \"Phase identification\nin electric power distribution systems by clustering of smart meter\ndata,\" in Machine Learning and Applications (ICMLA), 2016 15th IEEE\nInternational Conference on. IEEE, 2016, pp. 259\u2013265.", "W. Wang, N. Yu, and Z. Lu, \"Advanced metering infrastructure data\ndriven phase identification in smart grid,\" GREEN 2017 Forward, pp.\n16\u201323, 2017.\n[10] C. M. Bishop, Pattern Recognition and Machine Learning (Information\nScience and Statistics). Secaucus, NJ, USA: Springer-Verlag New York,\nInc., 2006.\n[11] Y. Le Borgne, \"Bias-variance trade-off characterization in a classification\nproblem: What differences with regression,\" Machine Learning Group,\nUniv. Libre de Bruxelles, Belgium, 2005.\n[12] K. Hajebi, Y. Abbasi-Yadkori, H. Shahbazi, and H. Zhang, \"Fast\napproximate nearest-neighbor search with k-nearest neighbor graph,\"\nin IJCAI Proceedings-International Joint Conference on Artificial\nIntelligence, vol. 22, no. 1, 2011, p. 1312.\n[13] W.-Y. Loh, \"Classification and regression tree methods,\" Encyclopedia\nof statistics in quality and reliability, 2008.\n[14] B. P. Roe, H.-J. Yang, J. Zhu, Y. Liu, I. Stancu, and G. McGregor,\n\"Boosted decision trees as an alternative to artificial neural networks\nfor particle identification,\" Nuclear Instruments and Methods in Physics\nResearch A, vol. 543, pp. 577\u2013584, May 2005.\n[15] S. Sonoda and N. Murata, \"Neural network with unbounded activation\nfunctions is universal approximator,\" ArXiv e-prints, May 2015.\n[16] D.-A. Clevert, T. Unterthiner, and S. Hochreiter, \"Fast and accurate deep\nnetwork learning by exponential linear units (ELUs),\" ArXiv e-prints,\nNov. 2015.\n[17] G. Klambauer, T. Unterthiner, A. Mayr, and S. Hochreiter,\n\"Self-normalizing neural networks,\" ArXiv e-prints, Jun. 2017.\n[18] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. MIT Press,\n2016, http://www.deeplearningbook.org.\n[19] J. Lampinen and A. Vehtari, \"Bayesian approach for neural\nnetworksreview and case studies,\" Neural networks, vol. 14, no. 3, pp.\n257\u2013274, 2001.\n[20] D. M. Blei, A. Kucukelbir, and J. D. McAuliffe, \"Variational inference:\nA review for statisticians,\" ArXiv e-prints, Jan. 2016.\n[21] R. Ranganath, S. Gerrish, and D. Blei, \"Black box variational inference,\"\nin Artificial Intelligence and Statistics, 2014, pp. 814\u2013822.\n[22] Y. Gal and Z. Ghahramani, \"Dropout as a bayesian approximation:\nRepresenting model uncertainty in deep learning,\" in International\nConference on Machine Learning, 2016, pp. 1050\u20131059.\n[23] H. Lin and J. Bilmes, \"How to select a good training-data subset for\ntranscription: Submodular active selection for sequences,\" Washington\nUniversity Seattle Dept. of Electrical Engineering, Tech. Rep., 2009.\n[24] U. Von Luxburg, \"A tutorial on spectral clustering,\" Statistics and\ncomputing, vol. 17, no. 4, pp. 395\u2013416, 2007.\n[25] A. Krause and D. Golovin, \"Submodular function maximization.\" 2014."]}

Power distribution circuits undergo frequent network topology changes that are often left undocumented. As a result, the documentation of a circuit’s connectivity becomes inaccurate with time. The lack of reliable circuit connectivity information is one of the biggest obstacles to model, monitor, and control modern distribution systems. To enhance the reliability and efficiency of electric power distribution systems, the circuit’s connectivity information must be updated periodically. This paper focuses on one critical component of a distribution circuit’s topology - the secondary transformer to phase association. This topology component describes the set of phase lines that feed power to a given secondary transformer (and therefore a given group of power consumers). Finding the documentation of this component is call Phase Identification, and is typically performed with physical measurements. These measurements can take time lengths on the order of several months, but with supervised learning, the time length can be reduced significantly. This paper compares several such methods applied to Phase Identification for a large range of real distribution circuits, describes a method of training data selection, describes preprocessing steps unique to the Phase Identification problem, and ultimately describes a method which obtains high accuracy (> 96% in most cases, > 92% in the worst case) using only 5% of the measurements typically used for Phase Identification.