Construction Method of Structured Knowledge Network Based on Abdominal Ultrasound Text Data

Abstract

The invention provides a method for constructing a structured knowledge network of abdominal ultrasound text data. The method includes: using a word segmentation tool to perform text segmentation processing on Chinese abdominal ultrasound text data to obtain a text set of the abdominal ultrasound text data; calculating the similarity between different texts in the text set, and according to the similarity between the texts degree constructs a text similarity matrix, utilizes the text similarity matrix to realize the text clustering of described abdominal ultrasound text data, and locates and maps the short sentences in the abdominal ultrasound text data after text clustering; according to the positioning of the short sentences With the mapping result, the entity, attribute and value of the abdominal ultrasound text data are marked by an algorithm, and the structured knowledge network of the abdominal ultrasound text data is constructed based on the entity, attribute and value. The ultrasonic examination knowledge map established by the method of the present invention fully retains the ultrasonic examination knowledge, can be stored in a structured manner, and provides a data basis for higher-level intelligent diagnosis decision-making application scenarios.

Description

technical field [0001] The invention relates to the technical field of text structured processing, in particular to a method for constructing a structured knowledge network of abdominal ultrasound text data. Background technique [0002] As an important class of unstructured clinical documents, medical pathology texts are crucial to clinical diagnosis. In practical applications, it is necessary to perform structured processing on Chinese medical pathology text data. [0003] In recent years, domestic scholars have begun to pay attention to the text mining of electronic medical records. There is a research specifically for Chinese electronic medical record texts, and the word segmentation method based on existing word segmentation tools can achieve a precision of up to 78.06%; some research uses electronic medical record texts as the basis to mine the potential semantics of discharge records, but this research only focuses on Four treatment options have been evaluated, and ...

AI Technical Summary

Problems solved by technology

[0004] Ultrasound examination texts usually mainly involve the examination content of the liver, gallbladder, pancreas, spleen, and kidney. The examination records are generally about 200 to 300 words, and contain many special words and idioms, and the text complexity is relatively high

Based on the idea of ​​information extraction, some studies have structured the thyroid description text in ultrasound examination, but they have not fully covered the ultrasound examination text, nor have further established the relationship network of the structured content of the ultrasound examination text

Images