基于药物论坛中潜在不良反应与适应症的知识发现体系构建

目的：基于用户反馈信息，构建药物论坛中潜在不良反应与适应症的知识发现体系。方法：利用BIO序列标注方法和BiLSTM-CRF命名实体识别方法，从药物论坛的评论数据中识别不良反应与适应症的实体。结果：对不良反应与适应症进行过滤，得到26种药物共180种不良反应和26种适应症，并以此构建模型。结论：基于药物论坛构建的知识发现体系为挖掘中文领域内基于评论信息药物潜在不良反应与适应症可提供一种新的研究思路方案。

Establishment of knowledge discovery system for potential adverse reactions of drugs and indications for for their use in drug forum

Objective To establish a knowledge discovery system for potential adverse reactions of drugs and indications for their use in drug forum based on the feedback information of users. Methods The adverse reactions of drugs and indications for their use were identified from the review data in drug forum by BIO sequence labeling and BiLSTM-CRF named entity recognition. Results One hundred and eighty kinds of adverse reactions of drugs and 26 indications for their use were discovered by screening their adverse reactions and indications, based on which a model was established. Conclusion The knowledge discovery system for potential adverse reactions of drugs and inactions for their use in drug forum can provide a new tool for discovering the potential adverse reactions of drugs and indications for their use in Chinese papers based on their review data.