C. Nalmpantis, O. Krystalakos, and D. Vrakas. 2018. Energy profile representation in vector space. In Proceedings of the 10th Hellenic Conference on Artificial Intelligence (SETN ’18). Association for Computing Machinery, New York, NY, USA, Article 22, 1–5. DOI:https://doi.org/10.1145/3200947.3201050

Author(s): Christoforos Nalmpantis, Odysseas Krystalakos, and Dimitris Vrakas


Keywords: energy2vec, energy embeddings, word2vec, skip-gram


Abstract: Word2vec is a computationally efficient way to represent words in vector space. Word embeddings are now considered an integral part of many Natural Language Processing problems. Its applicability has been successfully extended to other sequential problems, mainly in the discrete space. In this paper, the applicability of word2vec in time-series is examined, specifically for energy related problems. A novel framework, called Energy2vec, is presented and future improvements are discussed.