A Novel Framework for Smart Agriculture Prediction using Machine Learning Techniques

Abstract

Agriculture plays a significant role in the financial system worldwide, demand in the agricultural system will rise with the ongoing growth of the human population. Agriculture -technology and precision farming, now also termed digital agriculture, have arisen as a novel technical field that use data centric approaches to have better yield. The data created in current agricultural operations given by a variety of sensors that allow a superior understanding of the operational environment (an interaction of dynamic soil, crop, and weather circumstances) and the operation itself (machinery data), leading to extra precise and quicker decision making. This research primarily aims to enhance agriculture technology. The research methodology is categorized into four phases, viz soil profile prediction, crop yield production forecasting, crop disease detection, groundwater and dam reservoir level forecasting. The experimental result shows that the proposed Modified Regression, Fuzzy C Means, Linear Discriminant Analysis and Artificial Neural Network model gives superior performance than the existing models. In this thesis, the first phase has taken up the issue of predicting the yield level by analyzing the soil characteristics for the particular soil. As gathering the information from the enormous volume of data is practically a complicated task in the recent circumstances. This phase has concentrated only Trichy district, and the soil data samples have been collected. A preprocessing element was premeditated to construct training and testing sets to use among the proposed and compared classifiers. The classifiers like Linear Regression, Simple Linear Regression, Additive Regression, and Regression by Discretization used for training and testing of the data sets. This phase has modified the Regression by Discretization classifier and applied. newlineThe second contribution of the work is Crop Yield Prediction Forecasting in Trichy District using Fuzzy C Means Algorithm and Multilayer Perceptron.