Quantile load forecasting on streaming data and adversarial transfer Bi-LSTM networks

Xiaodong Li; Jizhong Zhu; Hanjiang Dong; Xiang Lai; Miaomiao Zhou; Yanting Huang

doi:10.1117/12.2640694

24 October 2022 Quantile load forecasting on streaming data and adversarial transfer Bi-LSTM networks

Xiaodong Li, Jizhong Zhu, Hanjiang Dong, Xiang Lai, Miaomiao Zhou, Yanting Huang

Proceedings Volume 12289, International Conference on Intelligent Manufacturing and Industrial Automation (CIMIA 2022); 122890U (2022) https://doi.org/10.1117/12.2640694
Event: International Conference on Intelligent Manufacturing and Industrial Automation (CIMIA 2022), 2022, Kunming, China

Abstract

Streaming data refers to a dynamic data set that grows infinitely over time. And the streaming data set of load lacks historical data due to difficult data collection or short running time. Therefore, it is very difficult to ensure the accuracy of streaming data forecasting results. To solve the problem, this paper proposes a quantile prediction model for streaming data based on the adversarial transfer bidirectional long short-term memory (ATLSTM) networks. The model can be divided into two parts. In the pre-training model based on domain adversarial network, the adversarial mechanism is introduced based on the bidirectional long short-term memory (Bi-LSTM) architecture to train. In this paper, streaming data is used as the target domain data. And data with high similarity to streaming data is used as source domain data. By minimizing the prediction loss and maximizing the domain classification loss, the model extracts the common features of the target domain and the source domain. In the migration training model based on parameter fine-tuning, a simple model is constructed by importing part of the network of the pre-training model. On this basis, the optimization training is carried out through the data of the target domain to extract the deep features. And the quantile fitting can not only solve the situation of outliers in the data, but also obtain the load forecasting curve under different confidence conditions. The experimental results show that the accuracy, efficiency, and stability of the proposed model are better than the traditional deep learning model.

Citation Download Citation

Xiaodong Li, Jizhong Zhu, Hanjiang Dong, Xiang Lai, Miaomiao Zhou, and Yanting Huang "Quantile load forecasting on streaming data and adversarial transfer Bi-LSTM networks", Proc. SPIE 12289, International Conference on Intelligent Manufacturing and Industrial Automation (CIMIA 2022), 122890U (24 October 2022); https://doi.org/10.1117/12.2640694

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
9 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Data modeling

Feature extraction

Neural networks

Mathematical modeling

Control systems

Data conversion

Process modeling

Show All Keywords

Keywords/Phrases

Search In:

Publication Years