Hospital infection intelligent diagnosis method based on multi-model fusion and system

Abstract

The invention discloses a hospital infection intelligent diagnosis method based on multi-model fusion and a system. The method comprises the steps of acquiring multiple medical record data related with hospital infection; preprocessing the medical record data, and obtaining a plurality of discrete word lists which correspond with each packet of medical record data; dividing all word lists into a training set and a testing set according to a proportion; in the training set, obtaining an optimal characteristic set for aiming at different infection types; performing parameter adjustment on more than two base models, selecting an optimal parameter for obtaining more than two optimal base models, fusing all optimal base models for obtaining a diagnosis model; testing the diagnosis model by means of the testing set, and analyzing the performance of the diagnosis model. The hospital infection intelligent diagnosis method and the system settle problems of low accuracy and low reporting omission rate caused by a single diagnosis model and acquires the diagnosis model through multiple model fusion, thereby improving diagnosis accuracy and reducing reporting omission rate.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a multi-model-based intelligent diagnosis method and system for hospital infection. Background technique [0002] Nosocomial infection refers to the infection acquired by inpatients in the hospital. Usually, the diagnostic steps can be divided into two steps: first, determine whether the infection is based on clinical data, laboratory test results, and various professional diagnostic indicators; and then analyze whether the infection belongs to nosocomial infection. The diagnostic principles of nosocomial infection are divided into two categories. For infectious diseases with a clear incubation period, the calculation starts from the first day of admission, and the infection that occurs after the average incubation period is nosocomial infection. For those with an unclear incubation period, it is generally considered that the infection occurred after 48 hours after admissi...

AI Technical Summary

Problems solved by technology

[0008] The present invention provides a multi-model based intelligent diagnosis method and system for nosocomial infection, which is used to overcome the defects of low diagnosis rate and significant false negative rate caused by a single calculation model in the prior art, and process data through multi-model fusion, Improve the accuracy of diagnosis and reduce the false negative rate

Images