Read the full-text of this article for free

Title: Automatic completion algorithm for knowledge graphs in power market based on convolutional neural network

Authors: Zhenhua Cai

Addresses: School of Artificial Intelligence, Zhangzhou Institute of Technology, Zhangzhou, 363000, Fujian, China

Abstract: The application of knowledge graphs (KGs) in the electricity market industry is very limited, and their completeness urgently needs to be improved. Therefore, this paper uses convolutional neural networks (CNN) to study the automatic completion algorithm of professional KG in the electricity market. This paper improves the performance of the algorithm by supplementing existing KG, and after initialising power data using CNN algorithm, uses convolutional kernels to extract features from triple professional information in the electricity market. Through experiments, it can be found that the professional KG auto completion algorithm based on CNN algorithm has a recognition accuracy of over 90.5% in the electricity market, and the average recognition accuracy after 10 experiments is 91.5%. Compared with other KG completion algorithms, this algorithm has excellent performance. The CNN algorithm can effectively extract triplet information of electricity.

Keywords: power market; knowledge graphs automatic completion algorithm; CNN; convolutional neural network; ESI; entity similarity information; electricity information triple.

DOI: 10.1504/IJEP.2026.156123

International Journal of Environment and Pollution, 2026 Vol.76 No.8, pp.27 - 43

Received: 24 Sep 2025
Accepted: 27 Jan 2026
Published online: 04 Sep 2026 *