摘要
为了将实体识别技术应用于医疗信息系统,提取医学影像检查报告的特征数据,提出了一种基于BiLSTM-CRF模型的信息实体识别方法。构建医学影像检查报告的智能识别系统,实现部位、症状等关键内容的结构化解析,通过可用性评估来分析应用效果。该识别系统已投入使用,共处理了3446份胸部放射CT报告。实验结果表明其识别精度较高,智能提示功能提高了医生的满意度。由此可见命名实体识别方法有助于挖掘医疗文本的价值,在医疗大数据领域有着广阔的应用前景。
In order to apply entity recognition technology to medical information system to extract characteristic data of medical imaging examination report,an information entity recognition method based on BiLSTM-CRF model is proposed to establish an intelligent recognition system for medical imaging examination report,and realize the structural analysis of key contents such as location and symptoms.The application effect is analyzed through usability evaluation.With the application of the recognition system,a total of 3446 chest radiology CT reports are processed.The experimental results show that the recognition accuracy is high,and the intelligent prompt function improves the satisfaction of doctors.Named entity recognition method is helpful to mining the value of medical texts and has a broad application prospect in the field of medical big data.
作者
尤丽珏
尹远芳
YOU Lijue;YIN Yuanfang(Huadong Hospital,Shanghai 200040,China)
出处
《微型电脑应用》
2023年第10期134-137,共4页
Microcomputer Applications
