Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Forecasting Rice Production of Bangladesh Using Arima and Artificial Neural Network Models

Abira Sultana
Department of Statistics, University of Dhaka, Dhaka-1000, Bangladesh

Murshida Khanam
Department of Statistics, University of Dhaka, Dhaka-1000, Bangladesh

Abstract/ Executive Summary:

Forecasting behavior of Econometric and Machine Learning models has recently attracted much attention in the research sector. In this study an attempt has been made to compare the forecasting behavior of Autoregressive Integrated Moving Average (ARIMA) and Artificial Neural Network (ANN) using univariate time series data of annual rice production (1972 to 2013) of Bangladesh. Here, suitable ARIMA has been chosen from several selected ARIMA models with the help of AIC and BIC values. A simple ANN model using backpropagation algorithm with appropriate number of nodes or neurons in a single hidden layer, adjustable threshold value and learning rate, has been constructed. Based on the RMSE, MAE and MAPE values, the results showed that the estimated error of ANN is much higher than the estimated error of chosen ARIMA. So, according to this study, it can be said that the ARIMA model is better than ANN model for forecasting the rice production in Bangladesh.

Keywords: ARIMA, ANN, AIC, BIC, Nodes, Hidden Layer, Learning rate.

Location: Department of Statistics, University of Dhaka, Dhaka-1000, Bangladesh

Start Date: 00-00-1972

End Date: 00-00-2013

Program Area: Knowledge Management

Commodity: Modeling, Rice

Objectives:

This work differs the other mentioned researches in many ways. In this paper, the forecasting behavior of selected ARIMA and ANN models is compared using univariate time series data of annual rice production (1972 to 2013) and the best of these two models considering the values of RMSE, MAPE and MAE is selected for further forecasting of Rice production.

Methodology:

The two models that have been considered in the present study are ARIMA and ANN Models.

ARIMA Model ARIMA model popularly known as the Box-Jenkins (BJ) methodology which is the most general class of models for forecasting a time series data. The data can be made stationary by differencing or taking logarithm. A series that needs to be differenced to make stationary is called “integrated”.

A series is called ARIMA (p,d,q) if we need to difference the series d times to make it stationary before applying ARMA (p,q) model. Here, p denotes the number of autoregressive terms and q denotes the number of moving average terms.

ANN Model An ANN or connectionist system is computing system vaguely inspired by the biological neural networks that constitute animal brains. Such system learns to perform tasks by considering examples, generally without being programmed with any task specific rules.

An ANN is reinforcement on a group of linked units or nodes called artificial neurons like the neurons in a biological brain. Every association is like the synapse in a biological brain. This transmits a sign from one artificial neuron to another.

Components of ANN Several components are involved to construct an ANN. Some are: Neurons, Connections and Weights, Propagation function, Learning rule and Learning algorithm.

Stationarity Test for the Data Set Before starting the analysis of the study, we have tested the stationarity of the data set. The overall stationarity test is conducted using two tests, Graphical Analysis and Augmented Dickey-Fuller (ADF) test.

Evaluation of Forecasting Method There exist many statistics to evaluate the forecast error of any Time Series or Econometric model. Let us suppose that, Yt be the actual observation and ft is the fitted value. Then the following statistical measures can be considered to evaluate the forecasting.

Source of Information: Dhaka Univ. J. Sci. 68(2): 143- 147, 2020 (July)

Article Link (Online Journal):

Funding Source:

Budget:

Total Budget (Tk.) :

Achievement/Findings:

According to this study, ARIMA model has been recommended to forecast the time series annual data. Because in this study, ARIMA model gives better result comparing with the ANN model for forecasting the Rice production of Bangladesh.

In this paper, univariate data has been considered to compare ARIMA model and ANN model. It can be suggested to do further research considering bivariate data and make a comparison between ARIMA model and ANN model. It can be recommended to make comparison with other model such as VAR model, VECM model, Holt's smoothing techniques etc. Moreover, this work is a collaboration of artificial intelligence and econometric models. Works like this need to be executed more in the field of several researches.

Knowledge Management: Journal