Artificial neural network and regression models for predicting Fiji population

Author

Qiokata, Viliame ; Khan, M.G.M.

Author_Institution

Coll. of Found. Studies, Univ. of the South Pacific, Suva, Fiji

fYear

2014

Firstpage

1

Lastpage

7

Abstract

The paper compares Artificial Neural Network (ANN) model against traditional models in the modeling of population and external migration for Fiji population components during the years from 1986 to 2012. The performance of the various models used are based on the values of the various error functions such as the R-squared (R2), Root Square Mean Error (RSME), Mean Absolute Error (MAE), Standard Error of Regression (SER), Sum Squared Residual (SSR), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). Across yearly time series, ANN performed better than those traditional models, when comparing the various error functions used.

Keywords

neural nets; regression analysis; social sciences computing; ANN model; Fiji population prediction; MAE; MAPE; R-squared; R2; RSME; SER; SSR; artificial neural network; error functions; mean absolute error; mean absolute percentage error; regression models; root square mean error; standard error of regression; sum squared residual; Data models; Equations; Mathematical model; Predictive models; Sociology; Time series analysis; Artificial Neural Network Model Time Series; Error Functions;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Science and Engineering (APWC on CSE), 2014 Asia-Pacific World Congress on

Print_ISBN

978-1-4799-1955-0

Type

conf

DOI

10.1109/APWCCSE.2014.7053850

Filename

7053850