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

Predicting the Rice Production of Bangladesh by Machine Learning Technique

Shohel Mahmud*
Department of Statistics, Jahangirnagar University, Savar, Dhaka, Bangladesh

Abstract/ Executive Summary:

Bangladesh is an agricultural country and its economic condition largely depends on agriculture. The country produces much agricultural merchandise like rice, jute, wheat, onion, chili, banana, garlic, ginger, pulse and so on. However, rice (Oryza Sativa) is produced most widely all over the country. Moreover, the United States Department of Agriculture (USDA), 2017 estimates that the rice production of Bangladesh is 34.7 million metric tons for the period 2017–2018 and the position of Bangladesh is after China, India, and Indonesia. In addition, Bangladesh produces three types of rice which are Aus, Aman, and Boro. A huge portion of the population in Bangladesh immensely depends on rice as the main food. So, this paper attempts to predict the rice production of Bangladesh with the help of machine learning models like Artificial Neural Network (ANN). This paper considers a secondary data set of yearly rice production in Bangladesh over the period 1971–1972 to 2014–2015. This paper identifies the most suitable neural network model with architecture ANN 3*3*2*1 based on model selection criteria like MSE. Thus, this paper suggests an ANN model for envisaging the rice production of Bangladesh.

Keywords: Machine learning, Prediction, Rice prediction, Neural network

Location: Bangladesh

Start Date: 00-00-1970

End Date: 00-00-2015

Program Area: Socio-economic and Policy

Commodity: Rice

Objectives:

In this paper, we want to present an ANN model to predict the production of Aus, Aman, Boro in Bangladesh on the basis of some model selection criterion. Different activation functions were tested along with the different combinations of hidden layers to get a suitable model.

Methodology:

To set the model, secondary data of the rice production in Bangladesh over the period 1970–1971 to 2014–2015 published by the Bangladesh Bureau of Statistics (BBS) is used. Total 41 rice production data were analyzed, where three lag values were taken. 75% of data were used as training data to train the network. To predict the production of the next year of Aus, Aman, Boro and total yield of rice, R 3.4.1 version and R Studio 1.0.35 were used to set up the model of the rice production.ANNNeural Network (NN) often referred to as Artificial Neural Network to distinguish from the Biological Neural Networks. The key element of this illustration is the novel structure of the information processing system. It is composed of a large number of highly interconnected processing elements which are called neurons, working in unison to solve specific problems. ANNs like people, learn by example. An ANN is configured for a specific application, such as pattern recognition or data classification, through a learning process. Learning in biological systems involves adjustments to the synaptic connections that exist between the neurons. This is true of ANNs as well. The process of receiving and sending signals is done in a particular manner like a neuron receives signals from another neuron through dendrites. The neuron sends signals at spikes of electrical activity through a long thin stand known as an axon and an axon splits these signals through synapse and sends it to the other neurons. Similar to biological neuron (Figure 1 [6]), Artificial Neural Network also has neurons which are artificial and they also receive inputs from the other elements or other artificial neurons and then after the inputs are weighted and added, the result is then transformed by a transfer function into the output. In this study, the model was developed on the basis of the multilayer perception using a backpropagation algorithm with weights. Tangent hyperbolic is used as the activation function of these models. Model Selection Criteria Different model selection criteria have been used to select the appropriate model. Like AIC, BIC, MSE, MAPE. In this paper, I have considered the MSE to select the best-fitted model. MSE of an estimator measures the average of the squares of the errors or deviations-that is the difference between the estimator and what is estimated.

Source of Information: Journal of Agricultural Science and TechnologyISSN: 2278-2206 (Online), ISSN: 2349-3682 (Print)Volume 7, Issue 3

Article Link (Online Journal):

Funding Source:

Budget:

Total Budget (Tk.) :

Achievement/Findings:

Predicting the production of three types of rice in Bangladesh using ANNs has been the focus herein. Data from 1970–1971 period to 2014–2015 period were used to obtain the desire model. Models with the minimum MSE were selected. Though different forecasting methods are available, on the basis of the data and information availability ANN is used here to predict rice production. Furthermore, ANNs is more flexible and effective in case of correlation among the data set and it does not require as much assumption as the time series model.

Knowledge Management: Journal