[目的/意义]提出一个药物不良反应本体的半自动构建方法，构建的细粒度药物不良反应本体为利用社交媒体挖掘潜在的药物不良反应信号提供语义资源库。[方法/过程]首先，采用业务层次和语言层次相分离的设计理念，将用户在社交媒体中评论的药物不良反应表示成"对象要素-属性要素-描述概念"的形式。细粒度体现在社交媒体用户对药物同一不良反应描述概念表达的多样性上。然后，基于深度学习的思想，利用基于word2vec的描述概念候选词抽取算法自动地抽取出更多的描述概念候选词构建本体。[结果/结论]以糖尿病药物的建模实例表明，提出的细粒度药物不良反应本体的半自动构建方案，提高了本体构建的智能化水平，构建的细粒度药物不良反应本体为利用社交媒体挖掘潜在的药物不良反应信号提供语义资源库。
[Purpose/significance] The semi-automatic construction method for the adverse drug reaction ontology is proposed. The constructed fine-grained ADR ontology provides a semantic resource library for exploiting potential ADR signals by using social media. [Method/process] Firstly, based on the design concept that separates the business level and language level, this paper expressed the adverse drug reaction discussed in the network health community in the form of "object-attribute-description". The fine granularity is reflected in the diversity of describing the same adverse drug reaction. Then, based on the idea of deep learning, the word2vec-based description concept candidate word extraction algorithm is used to automatically extract more description concept to construct ontology. [Result/conclusion] The modeling example shows that the fine-grained ADR ontology construction scheme proposed in this paper can improve the efficiency and intelligence level of ontology construction. At the same time, the constructed fine-grained drug adverse reaction ontology provides a semantic resource library for exploiting potential ADR signals by using social media.
[1] 于跃.基于大数据挖掘的药品不良反应知识整合与利用研究[D].长春:吉林大学,2016.
[2] YANG M, KIANG M, SHANG W. Filtering big data from social media-Building an early warning system for adverse drug reactions[J]. Journal of biomedical informatics, 2015,54(C):230-240.
[3] CHEE B W,BERLIN R,SCHATZ B.Predicting adverse drug events from personal health messages[J]. AMIA symposium, 2011(4):217-226.
[4] VAN G K, DE G L, DE J, et al. Consumer adverse drug reaction reporting-a new step in pharmacovigilance?[J]. Drug safety, 2002, 26(4):211-217.
[5] VAN H F, VAN W C, PASSIER A, et al. Motives for reporting adverse drug reactions by patient-reporters in the Netherlands[J].Europe journal clinical pharmacology, 2010, 66(11):1143-1150.
[6] YAN P, CHEN H, ZENG D. Syndromic surveillance systems[J]. Annual review of information science and technology, 2010, 42(1):425-495.
[7] ABOU T M, ROSSARD C, CANTALOUBE L, et al. Analysis of patients' narratives posted on social media websites on benfluorex's (Mediator?) withdrawal in France[J].Journal of clinical pharmacy & therapeutics,2014,39(1):53-54.
[8] LEAMAN R, WOJTULEWICZ L, SULLIVAN R, et al. Towards internet-age pharmacovigilance. Extracting adverse drug reactions from user posts to health-related social networks[C]//Proceedings of the 2010 workshop on biomedical natural language processing. Stroudsburg:Association for Computational Linguistics, 2010:117-25.
[9] VAN H F, TALSMA A, VAN P E, et al. The proportion of patient reports of suspected ADRs to signal detection in the Netherlands:case-control study[J]. Pharmacoepidemiology drug & safety 2011, 20(3):286-291.
[10] CAI M C, XU Q, PAN Y J, et al. ADReCS:an ontology database for aiding standardization and hierarchical classification of adverse drug reaction terms[J].Nucleic acids research, 2015(43):D907-D913.
[11] SOUVIGNET J, DECLERCK G, ASFARI H, et al. OntoADR a semantic resource describing adverse drug reactions to support searching, coding, and information retrieval[J]. Journal of biomedical informatics, 2016, 10(63):100-107.
[12] HE Y Q, SARNTIVIJAI S, LIN Y. OAE:The ontology of adverse events[J]. Journal of biomedical semantics, 2014, 5(1):1-13.
[13] 李梅. 我国心血管药物不良反应中英文本体构建与知识发现研究[D].长春:吉林大学,2017.
[14] 李梅,曹玉莹,张华吉, 等. 基于本体的国内抗感染药物不良反应报告分析[J]. 药物流行病学杂志,2017(2):115-119.
[15] 张佶.传统银行的转型实战[EB/OL].[2018-03-11].https://www.geekbang.org/.
[16] 王红滨,刘大昕,王念滨. 基于遗传算法和种子概念的本体概念提取算法[J].系统工程与电子技术,2010,32(11):2465-2469.
[17] 闭炳华. 基于word2vec的数字图书馆本体构建技术研究[J].现代电子技术,2016,39(15):90-94.
[18] 于娟,党延忠. 领域特征词的提取方法研究[J].情报学报,2009,28(3):368-373.
