A Method for Medical Diagnosis Based on Deep Convolutional Neural Networks

Abstract

The invention discloses a method for medical diagnosis based on a deep convolutional neural network. The method includes: obtaining a word vector matrix corresponding to an electronic medical record to be diagnosed; inputting the word vector matrix corresponding to the electronic medical record to be diagnosed into a pre-built deep volume In the product neural network model, the feature vector of the electronic medical record to be diagnosed is obtained; the feature vector of the electronic medical record to be diagnosed is classified by a classifier, and the prevalence probability of each disease corresponding to the electronic medical record to be diagnosed is obtained. This method applies the convolutional neural network to the semantic understanding of medical electronic medical record text and assists medical diagnosis, which can effectively overcome the defects of the method based on rule extraction and matching.

Description

technical field [0001] The invention relates to the technical field of medical informatization, in particular to a method for medical diagnosis based on a deep convolutional neural network. Background technique [0002] Clinical decision support system (CDSS) refers to the use of relevant and systematic clinical knowledge and basic patient information and condition information to strengthen medical-related decision-making and actions, and improve medical quality and medical service levels. CDSS is an important means to improve medical quality, and its fundamental purpose is to evaluate and improve medical quality, reduce medical errors, and control medical expenses. [0003] At present, the vast majority of CDSS in the world are implemented based on rule extraction and matching. The following technical solutions are adopted: based on clinical databases, logically related knowledge points are established by collecting, sorting, classifying, filtering, and processing informati...

AI Technical Summary

Problems solved by technology

This method can effectively solve the problem of "semantic gap" and effectively eliminate the impact of electronic medical records such as words with different synonyms in writing; Constructing corresponding rules and matching algorithms can build a unified model framework, and then use the historical data of each disease to train the model, so as to achieve the effect of prediagnosing multiple diseases with only one model, which is very suitable for management and maintenance, and can be expanded It is also very strong; this method does not need to design rules and features manually, all the features and rules learned by the model come from a large amount of clinical historical data, and it is completely guided by clinical historical data to make clinical decisions. Very strong practical guiding significance

Images