Title: Entity extraction based on the combination of information entropy and TF-IDF

Authors: Hankiz Yilahun; Askar Hamdulla

Addresses: Xinjiang Key Laboratory of Multilingual Information Technology, School of Information Science and Engineering, Xinjiang University, Urumqi, 830046, China ' Xinjiang Key Laboratory of Multilingual Information Technology, School of Information Science and Engineering, Xinjiang University, Urumqi, 830046, China

Abstract: Traditional knowledge graph entity extraction methods require expert knowledge and a large number of artificial features. Furthermore, deficiencies exist in the accuracy and efficiency of keyword extraction based on methods such as TF-IDF. Thus, this study proposes a Chinese entity extraction method based on the combination of information entropy and TF-IDF. First, the text is preprocessed, which involves operations such as sentence segmentation, word segmentation, removal of stop words, and POS tagging, to detect keywords based on POS. Secondly, the word frequency is analysed to determine feature word weight, and the TF-IDF algorithm is used to compare the importance of keywords. Finally, information entropy is used to improve the TF-IDF algorithm to provide entity knowledge for the construction of the knowledge graph. The entity extraction method and optimisation scheme proposed in this study can help users extract domain entities and provide better entity resources for the construction of knowledge graphs.

Keywords: entity extraction; improved TF-IDF; information entropy; knowledge graph.

DOI: 10.1504/IJRIS.2023.128371

International Journal of Reasoning-based Intelligent Systems, 2023 Vol.15 No.1, pp.71 - 78

Received: 11 Oct 2021
Accepted: 16 Mar 2022
Published online: 18 Jan 2023 *