
Defect Text Mining Technique and Application in Power Grid Based on the modified Semantic Framework
Author(s) -
Tsung Han Yang,
Y. H. Wang,
Han Zhao,
Yongfeng Liang,
Jinzhu Gao,
Xiaodong Ji
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/701/1/012037
Subject(s) - computer science , transformer , sentence , process (computing) , ontology , semantic grid , natural language processing , semantic computing , reliability (semiconductor) , data mining , word (group theory) , grid , power grid , information retrieval , artificial intelligence , power (physics) , semantic web , engineering , physics , linguistics , geometry , mathematics , epistemology , quantum mechanics , voltage , electrical engineering , operating system , philosophy
There are mass data that contain important defect texts in the power grid enterprise, and they contain important reliability information. And the efficiency is very low to mine the exact information about the texts especially when the texts are in Chinese. Thus, the defect text mining technique based on the modified semantic framework is proposed. All texts are translated into English and use the text mining model based on the modified semantic framework, the defect texts are divided into a fixed pattern and the digital information can be extracted accurately. Take the transformer as an example, the first step is to establish the ontology dictionary and to separate the sentence and extract the texts’ features. Then, the modified power semantic framework and the semantic slots are defined, and the slots filling method and the semantic framework construction process are discussed, which can automatically perfect the ontology dictionary by merging the word series. Finally, the researches of defect text mining results of statistical reliability are studied, and the results show that the proposed model and method is feasible and effective when applied to automatic classification and statistics of grid defect.