Mathematics Applications for Agricultural Development

Tong, Hoh Le (2021) Mathematics Applications for Agricultural Development. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Tong Hoh Le.pdf Restricted to Registered users only Download (2MB)

Abstract

Rice and maize are one of the major cash crops in India. This study proposes a model that will enable us to measure how the crop responds to climate variables and the area of land. In India, being the oldest form of raising economy, agriculture is a field that has not yet been blessed by modern technology or data analysis. This study is made by analyzing the dataset of certain vital parameters for rice production and maize production, such as Temperature, and Area using Regression techniques to find out which regression models is the best predictive model. In addition, this study also wishes to use image processing methods such as Convolutional Neural Network and Fourier Transform techniques to detect the plant diseases and extract the feature. Based on the regression results, the best regression model to predict the rice production is the rice full model because it is the highest adjusted R square and meets all assumptions. Besides that, the best regression model to predict the maize production is the maize reduced model 2 since the maize full model has an assumption that does not meet the requirements of VIF, therefore the maize reduced model 2 becomes the most suitable model to predict the maize production. Apart from that, the result of the comparison of Convolutional Neural Network models had shown that the performance of CNN model with applying transfer learning in recognizing feature disease is much accurate and after applying Fourier Transform technique in image pre-processing which will become more easy detection of abrupt changes in images