Method of using semantic recognition for automatic coding conversion

Abstract

The invention provides a method of using semantic recognition for automatic coding conversion. The method includes: collecting original coding and a sample data set of a corresponding relation described by diagnosis; preprocessing data in a diagnosis text historical library according to medical rules, and performing word segmentation; building a synonym library for data in a diagnosis text training set, and performing preprocessing and word segmentation according to medical rules; calculating a document IDF weight value corresponding to each word in the diagnosis text historical library; performing word segmentation on each diagnosis text record in the diagnosis text training set to generate a training set TF-IDF matrix formed by text conversion; performing word segmentation on a to-be-converted diagnosis text record to acquire a word vector, comparing the word vector with the training set TF-IDF matrix, and finding disease coding corresponding to a most similar result in the training set through a cosine similarity formula. By the method, coding represented by text description can be converted automatically according to diagnosis text description written for a patient by a doctor.

Description

technical field [0001] The invention relates to the fields of medicine and computer application technology, in particular to a method for automatic disease code conversion using semantic recognition. Background technique [0002] The International Classification of Diseases (ICD) is a system that classifies diseases according to certain characteristics of diseases and expresses them by coding methods. It is the carrier for recording diagnostic information and the basis for medical data analysis, disease diagnosis grouping, performance evaluation, and medical insurance cost control. In the practice of medical institutions, the codes are refined differently based on the characteristics of the disease (ICD-10 clinical version), and different hospitals describe the diagnosis of the same disease differently. For example, "blepharospasm" in GB-ICD-10 has the same diagnostic meaning as "blepharospasm" in BJ-ICD-10, the former is coded as "G24.500", while the latter is coded as "G2...

AI Technical Summary

Problems solved by technology

The former method cannot be recognized when there is a slight deviation in the text, while the latter method has a large workload and requires high personnel ability and quality, making it difficult to ensure accurate conversion

Images