Font Size: a A A

Entity Relation Extraction Of Dermatosis Based On Dependency Syntax Analysis

Posted on:2020-11-09Degree:MasterType:Thesis
Country:ChinaCandidate:S JiangFull Text:PDF
GTID:2415330578958610Subject:Chinese Language and Literature
Abstract/Summary:PDF Full Text Request
This paper focuses on the construction of multi-view Chinese tree base and extraction of entity relations on the basis of linguistics.The main theoretical framework is text mining based on dependency syntax.The innovation of this paper is to combine dependency syntax tree base with specific professional knowledge,and then to name entity on the tagging platform of multi-view Chinese tree base.Selection and summary of named entity relations are processed,and then dependency syntactic models are extracted for various named entity relationships to show the semantic relationship between entities in dermatology.Combining theory with practice in order to expand corpus,enrich corpus scale and further expand the application of Chinese multi-view tree database,linguistics also contributes more to natural language processing,so as to provide a template and try more for computer to extract named entity relations in other related fields such as electronic medical records.In recent years,with the progress and maturity of natural language processing technology,a large number of unstructured text data involving various fields,such as dermatosis,textbook Monographs in primary and secondary schools,electronic medical records and so on,can try to carry out structured data processing to facilitate text mining.At present,most medical texts are not structured information data,including many complex unstructured textual data described by natural language.Although the information data of text type can describe concepts and events in detail and accurately,the large and unstructured textual data also add burden and system to search,statistical analysis and so on.Obstacles may have a negative impact on the quality of data analysis results.In order to make full use of existing linguistic theories and mature analytical tools to analyze and mine medical texts effectively,so as to improve the value of medical data,structured processing of unstructured text data has become a problem that many scholars focus on and need to find better solutions.Therefore,natural language processing,information extraction and other related technologies will play an important role in the analysis and mining of medical texts.Natural language processing(NLP)is applied to medical texts,also known as medical language processing(MLP).Its basic research includes word segmentation,part-of-speech tagging,parsing,named entity recognition,entity relationship extraction,and co-referential resolution.Deep-processing corpus(Treebank)can be used to study various grammatical phenomena and the overall characteristics of the language.This paper is an entity relationship extraction work on the platform of Multi-view Chinese Treebank.This paper is divided into five parts.The main contents of each part are as follows:The first part is introduction.This paper mainly introduces the research background,theoretical basis,research status,research methods and research significance.The second part is the multi-view Chinese Treebank.This paper mainly introduces the construction of multi-view Chinese Treebank,including the source of corpus,annotation system,annotation model and annotation platform.The third part is named entity and entity relationship.This paper mainly introduces the selection of entity relationship and the classification of entity relationship.The fourth part is entity relationship extraction.This paper mainly introduces the work related to entity relationship extraction.The last part,conclusions and prospects.This paper summarizes the main research contents and achievements of this paper,summarizes the significance of natural language processing and language research,analyses the shortcomings of this paper and makes a preliminary plan for the next step.
Keywords/Search Tags:Skin disease, Dependency parsing, Named entity, Relation ectraction, Corpus construction
PDF Full Text Request
Related items