Crop Recommendation Application using Ensemble Classifiers

Abstract

The practice of raising cattle and plants is known as agriculture. It necessitates the preparation of plant and animal products and distribution of them to markets for human consumption. Agriculture plays a major role in the world’s food and textile production. Wool, cotton, and leather are products of agriculture. Paper and lumber for building are additional products of agriculture. The agricultural methods used and the commodities produced can vary between different locations. The challenge for farmers is make to right choice of the crop in light of the current weather and soil nutrient levels. The project’s primary goal is to develop a reliable model that provides precise predictions of crop sustainability for a specific soil type and set of weather circumstances. In an attempt to eliminate loss for the farmers, the model provides a model that recommends the best local crop. The following factors are taken into account when building the model: nitrogen, potassium, phosphorus, temperature, air humidity, soil pH, and annual rainfall. Considering the data gathered from previous years, the model assists in choosing the type of crop that must be grown. The model is trained using an ensemble learning strategy that incorporates Gaussian Naive Bayes, Logistic Regression, and Support Vector Machine (SVM), which has acquired an accuracy of 99.31 for the model. These algorithms are evaluated at various levels and used as comparative research and analysis to support the task.