Construction Method And Implementation Of Neurodegenerative Disease Knowledge Base

The role of medical knowledge base in the field of modern biomedicine is becoming more and more important.In the era of big data,more hidden knowledge can be unearthed from the knowledge base to serve the fields of smart medical care.Especially for neurodegenerative diseases,the traditional model of new drug development and treatment models have fallen into a bottleneck period.Relevant entities are identified from the massive biomedical literature,and deep learning,relationship extraction,natural language processing and other technologies are used to construct the neural network of each entity and relationship.The knowledge base of degenerative diseases can assist drug researchers in new drug research and development,disease mechanism research and prevention and treatment strategies,etc.It has practical and practical significance for research in the field of biomedicine,among which efficient identification of biomedical entities from natural texts is a follow-up relationship The basic work of extraction and knowledge graph construction tasks.This article mainly includes the following two aspects of research:(1)This paper innovatively proposes an entity recognition algorithm based on capsule network to be applied to the field of biomedical entity recognition.This method takes the vector expression of the target word and surrounding words as input,and outputs the labels of the words before and after the target word,and obtains the label dependency to determine the label sequence.This method simplifies the structure of the traditional entity recognition algorithm based on recurrent neural network.It does not require a lot of manpower to add complex features.The model can automatically learn useful features.Even with large differences in corpus input,it can still achieve good performance.,The model can be migrated with a strong ability.Finally,through comparative verification,it is verified that the model performs well in the named entity recognition task of NCBI,CDR and CHEMDNER corpus.(2)In this paper,the above algorithm is applied to the construction of the neurodegenerative disease knowledge base,and the related symptoms,drugs,compounds,natural products,genes,peptides and other entities are identified from the biomedical literature related to neurodegenerative diseases and visualized in the form of a knowledge map Show.In addition,it also includes functions such as electronic medical record entry,rapid entity recognition of articles,and manual maintenance of entity equivalence.It supports multiple ways to enrich the knowledge base of neurodegenerative diseases.