Estimating The Number Of Daily Patient Applications By Using Artificial Neural Networks

İbrahim Güngör, Güngör (2012) Estimating The Number Of Daily Patient Applications By Using Artificial Neural Networks. In: 3rd International Symposium on Sustainable Development, May 31 - June 01 2012, Sarajevo.

[img]
Preview
Text
12. Estimating The Number Of Daily Patient Applications By Using Artificial Neural.pdf

*- Download (522kB) | Preview

Abstract

This study is aiming at estimating the patient volumes of hospitals by using artificial neural networks. In order to train the artificial neural network models in this study, historical patient applications data from a Turkish hospital were used. All patient applications counted as daily numbers during three years and dependent variable of our study (patient_count) is derived. A different approach used in this study and instead of a single independent variable (which is time), four different time periods were used as input variables of the artificial neural network models. These input variables were day of month, day of week, month, and year. Several artificial neural network models have been generated and compared with each other by their predictive performance measures. The best predictive artificial neural network architecture has an estimation accuracy of 94.22 percent. This artificial neural network model has an input layer with four neurons, an output layer with one neuron, and only one hidden layer with nineteen neurons. The arithmetic mean of patient application in a day is 755.93 (S.d.=486.60). Mean error of the artificial neural network model is -0.047 and mean absolute error is 105.64. The linear correlation between the actual values and the predicted values of the number of patients is 0.918. Keywords: artificial neural networks, decision support systems, modeling, estimation, hospital management.

Item Type: Conference or Workshop Item (Paper)
Subjects: H Social Sciences > H Social Sciences (General)
Divisions: Faculty of Economics > Management Department
Depositing User: Users 173 not found.
Date Deposited: 17 Oct 2012 12:42
Last Modified: 17 Oct 2012 12:42
URI: http://eprints.ibu.edu.ba/id/eprint/1105

Actions (login required)

View Item View Item