Author(s): Sousa, Regina ; Oliveira, Daniela Sofia Rijo ; Durães, Dalila ; Neto, Cristiana ; Machado, José Manuel
Date: 2022
Persistent ID: https://hdl.handle.net/1822/90818
Origin: RepositóriUM - Universidade do Minho
Subject(s): Emergency health department; Natural language processing; Recommendation systems; Text mining
The operational management of an emergency department (ED) requires more attention from hospital administration since it can have a global impact on the institution’s management, increasing the probability of adverse events and worsening hospital expenses. Effective management of an ED potentially results in fewer hospitalisations after an ED admission. The purpose of the present study is to perform a multi-class prediction based on: a) structured data and unstructured data in an ED episode; and b) unstructured data generated during the inpatient event, just after the ED episode. The designed prediction model will lay the foundation for an ED Decision Support System based on symptoms and principal diagnoses.
