JISE


  [ 1 ] [ 2 ] [ 3 ] [ 4 ] [ 5 ] [ 6 ] [ 7 ] [ 8 ] [ 9 ] [ 10 ] [ 11 ] [ 12 ] [ 13 ] [ 14 ] [ 15 ]


Journal of Information Science and Engineering, Vol. 36 No. 6, pp. 1167-1177


Forecasting Based on Some Statistical and Machine Learning Methods


AZHARI A. ELHAG1 AND ABDULLAH M. ALMARASHI2
1Mathematics and Statistics Department
Taif University
Taif, 21974 Saudi Arabia

2Statistics Department
King Abdulaziz University
Jeddah, 21589 Saudi Arabia
E-mail: azhri_elhag@hotmail.com


Forecasting consists basically of using data to predict the value of the attributes to promote micro- and macro-level decision making. There are many methods to do prediction extending from complexity and data requirement. In this paper, we present the method of an autoregressive integrated moving average (ARIMA), multilayer perceptron artificial neural network (ANN) model and decision tree (DT) method to forecast time-series data, also we use different methods to measure the accuracy of the forecasting of the patient dying after having Ebola virus in the Republic of Liberia over the period of 25 March 2014 to 13 April 2016. The data source is from World Health Organization (WHO).


Keywords: time series, modeling, deep learning, multilayer perceptron, forecasting

  Retrieve PDF document (JISE_202006_02.pdf)