基于语义框架的电网缺陷文本挖掘技术及其应用

发布时间：2018-03-04 07:14

本文选题：文本挖掘　切入点：语义框架　出处：《电网技术》2017年02期 　论文类型：期刊论文

【摘要】：电网企业拥有大量蕴含着重要可靠性信息的设备缺陷文本,依靠人工进行挖掘不仅效率低而且准确性因人而异。以变压器缺陷文本为研究对象,通过分析文本的特点,建立了基于语义框架的电网缺陷文本挖掘模型,解决了缺陷文本句子成分难以划分、数字量无法精确提取等问题,为电网领域的非结构化数据挖掘提供了新技术。首先在建立本体词库基础上,对缺陷文本进行分词、词汇特征提取等预处理;然后定义了电力语义框架与语义槽,提出了槽填充和语义框架构建流程,并通过词串合并实现了本体字典自动完善;最后对缺陷文本挖掘结果在可靠性统计中的应用进行了研究。算例表明,所提出的挖掘技术应用于电网缺陷自动分类与统计中,具有可行性和有效性。
[Abstract]:Power grid enterprises have a large number of equipment defect texts containing important reliability information. Relying on manual mining is not only inefficient but also different in accuracy. Taking transformer defect text as the research object, this paper analyzes the characteristics of the text. A model of grid defect text mining based on semantic framework is established, which solves the problems such as the difficulty of dividing the sentence components of defective text, the difficulty of extracting the exact number of defects, etc. This paper provides a new technology for unstructured data mining in power grid. Firstly, based on the establishment of ontology lexicon, word segmentation and lexical feature extraction of defect text are preprocessed, and then the power semantic framework and semantic slot are defined. The flow of slot filling and semantic frame construction is proposed, and the ontology dictionary is automatically perfected by string merging. Finally, the application of defect text mining results in reliability statistics is studied. The proposed mining technology is feasible and effective in the automatic classification and statistics of power grid defects.
【作者单位】：浙江大学电气工程学院;国网浙江金华供电公司;
【分类号】：TM732;TP391.1

【相似文献】