CROP YEILD PREDICTION USING ML

Abstract

In India's economy, agriculture is by far the most significant
industry, and it has the greatest impact on the country's gross
domestic product (GDP). An estimated 50 percent of the
country's workforce is employed in the industry, which accounts
for around 18 percent of the country's Gross Domestic Product
(GDP). People in India have been engaged in agricul ture fo r a
long time, but the results have never been satisfactory owing to a
variety of variables that influence crop productivity at d i fferent
times of the year in different regions. A high agricultural
production is required to meet the demands of the world's
approximately 1.2 billion people in order to ensure that they a re
met. All of the variables that influence crop output are d irectly
related to soil type, precipitation, seed quality, and the existence
or lack of technical infrastructure, to name a few. To meet the
increased demand, new technologies are required, and fa rmers
must use their resources effectively by embracing new technology
rather than relying on inefficient farming practises. The purpose
of this project is to demonstrate how to develop a crop production
forecast system using Data Mining methods. The dataset
pertaining to agriculture was the topic of the investigation.
Several classifiers, including the J48, LWL, LAD Tree, and IBK
are used to forecast it. The performance of each classifier is
evaluated by comparing its performance to the others using the
WEKA tools for enhancing Python with machine learning
performance (python with machine learning). In order to
evaluate total performance, it is necessary to include Accuracy
factors such as linear regression, as well as the accuracy of
Random forest and KNN classifiers, were employed in this study,
and one of them was the accuracy of linear regression. The
overall performance of the classifiers is then assessed by
comparing their Root Mean Squared Error (RMSE), Mean
Absolute Error (MAE), and Relative Absolute Error (RAE)
values to the values of Root Mean Squared Error (RMSE)
obtained from the training data (RAE). As a result, the technique
will perform more correctly as the number of errors lowers.
Classifiers are evaluated based on how well they perform in
classification by making comparisons with one another.