Advanced Soil Fertility Analysis and Crop Recommendation using Machine Learning
A. Jhansi Swetha, G. Kalyani, B. Kirananjali
Abstract
Farming is the foundation of any country's monetary design and has a basic impact on the Indian economy. Farmers may utilize technology to help them decide what to cultivate and how to grow it. It aids in crop selection and improves productivity for farmers. An effective agriculture depends on healthy soil. Soils are of various types, each with its unique composition of minerals, organic matter, and properties that allow for the cultivation of various crops. Crops are advised based on soil, climatic conditions, humidity, and other elements to increase agricultural production. To forecast soil fertility and suggest the best crop, four different types of classifiers: Artificial Neural Network, Decision Tree, Random Forest, and K-Nearest Neighbors, have been used. The accuracy of the various algorithms was compared. The results show that Random Forest (RF), which has an accuracy of 98.63 % in the crop dataset and 92.61 percent in the soil dataset, is the most effective algorithm out of those applied on the datasets. Using MIT App Inventor, a mobile application is developed for this model. It helps farmers to choose the correct crop and increase productivity.