基于半监督和深度学习的生物实体关系抽取-计算机应用技术专业论文.docxVIP

Biomedical Biomedical Entity Relation Extraction Based on Semi—supervised Learning and Deep Learning Ab stract In recent years，with the rapid growth of biomedical literature，the technology of information extraction(IE)in the biological literature has been studied extensively，Until now， most of the works in information extraction for biomedical literature are related to relation extraction．In the biomedical domain，relation extraction mainly focuses on recognizing the biomedical entities name(proteins，drugs，diseases，genes，etc．)and extracting the semantic relations between entities．This thesis focuses Oil studying the relations between disease—symptom，symptom-therapeutic substance and protein-protein，and proposes a semi—supervised learning and deep learning methods to solve the problem that are lack of 1abeled data and rl-latqual feature construction in entities relations extraction．respectively． To SOlve the problem of the 1ack of labeled data in extracting relations among disease and symptom，and symptom and therapeutic substance，this thesis proposes two semi—supervised learning algorithms，Co—Training and Tri—Training，to construct the disease—symptom model and symptom—therapeutic substance model．In the training process，the feature kernel，graph kemel and tree kernel are used as input views of Co．Training and Tri．Training methods．In the Tri—Training method，we use ensemble leaming to integrate several classifiers．Experimental results show that Co--Training andDi·-Training algorithms Can both utilize the unlabeled data along with a few labeled examples to improve the classification performance．In addition，the performance of Tri-Training outperforms Co·Training in the experiment． Using semi··supervised learning methods for relation extraction of the disease··symptom and symptom—therapeutic substance，requires large scale of manual features，the quality ofthese features have direct impact on the experimental results．Moreover，the construction of a large number of features is