Automatic grouping method for DRGs based on convolutional neural network

Abstract

The invention discloses an automatic grouping method for DRGs based on a convolutional neural network. The method comprises the following steps: collecting case data and grouping the data according tomain diagnosis categories and related grouping modes for core disease diagnosis; digitally encoding the data; constructing a shallow convolutional neural network model, clustering feature vectors extracted by the convolutional network by using a k-means clustering method to obtain k category labels, and performing iterative training through a combination of the category labels and a classifier supervision network; and after the model training is finished, performing data grouping application. By utilizing the method provided by the invention, the defects of manual feature selection and additional data annotation for newly-added grouping categories are avoided, and automatic learning grouping can be carried out on data which is fuzzy and difficult in grouping.

Description

technical field [0001] The invention belongs to the field of computer medical technology, in particular to an automatic grouping method of DRGs based on a convolutional neural network. Background technique [0002] With the aging of the population and the development of new science and technology, the post-payment system of the medical insurance fund is likely to stimulate excessive medical services, and the pre-payment system is likely to cause deficiencies such as prevarication of serious patients and reduction of medical services. In many areas, medical insurance funds are at risk of insufficient funds. [0003] DRGs (Diagnosis Related Groups, Disease Diagnosis Related Groups) is a case combination method, which mainly groups cases according to the principle of similar clinical process and similar cost consumption. Pay according to the diseases in different groups, and provide targeted treatment to avoid the waste of medical resources. However, due to the unbalanced eco...

AI Technical Summary

Problems solved by technology

[0006] However, the grouping of certain diseases in various regions may be controversial, and there may be different groupings in conventional methods. Therefore, it is urgent to design a method that can integrate various actual information to classify difficult categories.

Images