A Summary of Research on Deep Learning in Time Series Learning Methods

Rong Cao

《International Core Journal of Engineering》 6卷9期 (2020/09) Pp. 201-208

https://doi.org/10.6919/ICJE.202009_6(9).0024

摘要

In recent years, deep learning has developed rapidly, and end-to-end applications in the field of time series analysis are becoming more and more mature. Therefore, this article discusses the time series learning method based on deep learning, and summarizes the latest deep learning methods for task representation, prediction, classification, and anomaly detection of time series from the aspects of application, architecture, and ideas. The in-depth study of learning solutions and future development trends provide a reference.

關鍵字

Deep learning ； Time series prediction ； Time series classification

參考文獻

King S Y, Hwang J N. Neural network architectures for robotic applications[J]. IEEE Transactions on Robotics & Automation, 1989, 5(5):641-657.

Graves, A., et al.: Supervised sequence labelling with recurrent neural networks, vol. 385. Springer (2012)

Akmeliawati R, Ooi P L, Kuang Y C. Real-Time Malaysian Sign Language Translation using Colour Segmentation and Neural Network[C]// Instrumentation & Measurement Technology Conference. IEEE, 2007.

Lecun Y L, Bottou L, Bengio Y, et al. Gradient-Based Learning Applied to Document Recognition[J]. Proceedings of the IEEE, 1998, 86(11):2278-2324.

Qing X, Niu Y. Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM[J]. Energy, 2018, 148: 461-468.