| With the continuous development of hospital informatization,the use of electronic medical record systems has become more common,which has accumulated a large amount of medical data resources.The history of present illness in the discharge summary is one of the important components of these medical data resources.It records the health status of inpatients and the treatment process,which contains a wealth of medical knowledge.However,the present illness history in discharge summary is often recorded in narrative medical text,which is an unstructured text format.It is difficult to directly apply machine learning or deep learning models for data mining and analysis,which hinders the reuse of medical data.Therefore,structuring the present illness history in discharge summary can explore the potential value of the data and make it better serve the medical data research work.In this paper,the history of present illness data in the discharge summary provided by a 3A grade hospital in Shanghai is researched,and an ontology-driven method of structuring the present illness history in the discharge summary is proposed.The main work of this paper is as follows:(1)Based on the professionalism and particularity of medical data,this paper builds a professional dictionary by collecting vocabulary from medical professional data and crawling data from health websites.Then use professional dictionary auxiliary word segmentation tools to improve the accuracy of word segmentation and ensure the quality of the ontology and the effect of text structuring.(2)This paper proposes a method to construct the ontology of the present illness history in the discharge summary.First,perform data preprocessing on the present illness history in the discharge summary,including data cleaning and text segmentation.Then,based on the data characteristics of the present illness history,a basic framework of ontology was constructed.In order to facilitate the expansion and maintenance of the ontology,the present illness history ontology was decomposed into four sub-ontologies to construct.The single sentence after the splitting of the description block is the basic processing unit,the entity and its attributes are identified by dictionary matching,and then filled into each sub-ontology.Finally,the sub-ontology is merged to complete the construction of the present illness history in the discharge summary ontology.(3)This paper proposes an ontology-driven method to structure the present illness history in the discharge summary.First,use regular expressions and a set of heuristic rules to identify the time in the present illness history.Then the present illness history is segmented according to the time segmentation algorithm,and the text is segmented in each segment to obtain different description blocks.Next,the description block is split into single sentences,the single sentence and the corresponding sub-ontology are used as the basic processing unit.By obtaining the branch subtree path algorithm and generating the semantic subtree algorithm,the time axis semantic subtree corresponding to the discharge summary medical history is obtained.Finally,the semantic subtree stored in XML is converted into the structured data of the present illness history in the discharge summary stored in the relational file.(4)This paper designs and implements a structured visual system for the present illness history in the discharge summary.According to clinical needs,the structured method proposed in this paper is embedded into the system,and various functional modules of the system are designed and implemented,including structured functional modules,structured result visualization modules,and thesaurus management functional modules.To sum up,this paper first constructed the present illness history in the discharge summary ontology,and then proposed an ontology-driven structured method.Using comparative experiments to verify the effectiveness and accuracy of the method in this paper,and finally designed and implemented a structured visualization system for the present illness history in the discharge summary. |