The application of bp neural network in hospitalization expense research

Abstract

Data on hospitalization expense for coronary heart disease from 2009 to 2010 in one tertiary hospital were taken as examples, so as to construct fitting models for hospitalization expense based on BP neural network. Sensitivity analysis of influence factors was executed to evaluate the effect degrees of influence factors on hospitalization expense. The purpose of this study was to evaluate application value of BP neural network used in hospitalization expense study, explore the analyze and evaluation method fitting for the construction and character of hospitalization expense, use medical costs rationally and control irrationally increase of hospitalization expense. The results showed that the main influence factors of hospitalization expense were hospitalization days, treatment outcome, times of rescuing and age, the comprehensive influences of which were 0.83101, 0.76113, 0.73227 and 0.44537 respectively. © 2011 Springer-Verlag Berlin Heidelberg.