北京测绘2024，Vol.38Issue(12)：1670-1676,7.DOI:10.19580/j.cnki.1007-3000.2024.12.005
北京市不动产登记运维问题智能分类
Intelligent classification of operation and maintenance issues for real estate registration in Beijing
摘要
Abstract
To improve the efficiency of daily operation and maintenance for real estate registration in Beijing and address the issues of low efficiency and long response time in manual processing,this paper proposed an automatic classification method for operation and maintenance issues based on the bidirectional encoder representations from transformers(BERT).Firstly,the BERT model was utilized to extract contextual semantic features of the operation and maintenance issue texts.Secondly,global max pooling technology was applied to extract the key category features of the texts.Finally,the SoftMax function was used to calculate the probabilities of each category,and the category with the highest probability was selected as the classification result.Experimental results demonstrate that the macro precision(MP),macro recall(MR),and macro-average F1 score of the method proposed in this paper all exceed 93%,significantly surpassing common text classification techniques.This fully proves the effectiveness of the method and provides certain reference significance for constructing an intelligent operation and maintenance system for real estate registration.关键词
不动产/智能分类/预训练语言模型/基于变换器的双向编码器表示模型(BERT)/数据集构建Key words
real estate/intelligent classification/pre-trained language model/bidirectional encoder representation from transformer(BERT)/dataset construction分类
天文与地球科学引用本文复制引用
董承玮,李云汉,邢晨,肖曼丽,刘世凡..北京市不动产登记运维问题智能分类[J].北京测绘,2024,38(12):1670-1676,7.基金项目
国家自然科学基金(42274029) （42274029）
