A Comparison of ARIMA, Neural Network and Linear Regression Models for the Prediction of Infant Mortality Rate

Author

Purwanto ; Eswaran, Chikkannan ; Logeswaran, Rajasvaran

Author_Institution

Fac. of Comput. Sci., Dian Nuswantoro Univ., Semarang, Indonesia

fYear

2010

fDate

26-28 May 2010

Firstpage

34

Lastpage

39

Abstract

The aim of this paper is to compare the performances of ARIMA, Neural Network and Linear Regression models for the prediction of Infant Mortality Rate. The performance comparison is based on the Infant Mortality Rate data collected in Indonesia during the years 1995 – 2008. We compare the models using performance measures such as Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE). The results show that the Neural Network model with 6 input neurons, 10 hidden layer neurons and using hyperbolic tangent activation functions for the hidden and output layers is the best among the different models considered.

Keywords

Economic forecasting; Linear regression; Mathematical model; Medical services; Neural networks; Neurons; Pediatrics; Predictive models; Root mean square; Time series analysis; ARIMA; Infant Mortality Rate; Linear Regression; Mean Absolute Error; Mean Absolute Percentage Error; Neural Network; Root Mean Square Error;

fLanguage

English

Publisher

ieee

Conference_Titel

Mathematical/Analytical Modelling and Computer Simulation (AMS), 2010 Fourth Asia International Conference on

Conference_Location

Kota Kinabalu, Malaysia

Print_ISBN

978-1-4244-7196-6

Type

conf

DOI

10.1109/AMS.2010.20

Filename

5489680