NLPBD ECUSTNLPBD ECUSTnlpbd-ecustnlpbd-ecustTXTcc-byJun 21, 20172017-06-21
Dataset
ChineseSymptomNameRecognition
This dataset contains the training data for distant supervision from Baidu Baike used in the paper named Combining Distant Supervision with Syntax-based Rules for Chinese Symptom Names Recognition in Electronic Medical Records.
About
This dataset contains the training data for distant supervision from Baidu Baike used in the paper named Combining Distant Supervision with Syntax-based Rules for Chinese Symptom Names Recognition in Electronic Medical Records.
We also make a demo for our method to recognize Chinese symptom from texts. The demo URL is http://dsc.nlp-bigdatalab.org:8090/recognition.html
Download & API
Symptoms crawled from websitestxt
We crawled symptom names from the symptom list pages of 8 mainstream health information websites. They are:
* 39Health: http://jbk.39.net/bw_t2/
* 123ask: http://tag.120ask.com/zhengzhuang/
* Familydoctor: http://jbk.familydoctor.com.cn/
* Fh21: http://zzk.fh21.com.cn/
* jianke: http://www.jianke.com/zhengzhuang/
* PCbaby: http://health.pcbaby.com.cn/zz/
* qqyy: http://zz.qqyy.com/
* xywy: http://zzk.xywy.com/
The symptom dataset contains __14,707 symptom names__.
Training data for distant supervision from Baidu Baiketxt
We distantly annotated 120,000 sentences from Baidu Baike as the training data for our CRF model. Each symptom with its start index and end index was annotated after the sentence.
Sources & license
License
- cc-by