摘要
本文提出了基于命名实体识别的住院费用预测算法。该算法通过MEMM与CRF(最大熵马尔可夫模型与条件随机场)的命名实体识别方法在部分恶性肿瘤医疗文本数据描述中提取命名实体。实验结果表明在命名实体识别阶段所用测试集的平均准确率较为一般(56%~60%),单样本的最好准确率较高(84%~90%)。
This paper proposed a hospitalization expense prediction algorithm based on named entity recognition.The algorithm used the named entity recognition method of MEMM and CRF(Maximum Entropy Markov Model and Conditional Random Field)to extract named entities from the description of some malignant tumor medical text data.Experimental results showed that the average accuracy of the test set used in the named entity recognition stage was relatively average(56%-60%),and the best accuracy of a single sample was relatively high(84%-90%).
作者
杨丽静
唐俊
沈伟富
陈擎炜
徐煌
莫丽
YANG Li-jing;TANG Jun;SHEN Wei-fu;CHEN Qing-wei;XU Huang;MO Li(Hangzhou Healthcare Information Center;School of Computer and Computing Science,Zhejiang University City College;Zhejiang Chinese Medicine University)
出处
《医院管理论坛》
2020年第8期74-77,共4页
Hospital Management Forum
关键词
恶性肿瘤
信息提取
命名实体识别
最大熵马尔可夫模型
条件随机场
卷积神经网络
Malignant tumor
Information extraction
Named entity recognition
Maximum entropy Markov model
Conditional random field
Convolutional neural network
