Data-Driven Precision Agriculture for Crop Prediction and Fertilizer Recommendation Using Machine Learning

Abstract

Crop prediction and fertilizer recommendation are essential for optimizing agricultural practices, a crucial concern for the agricultural sector. However, accurate crop and fertilizer prediction has long been challenging, requiring innovative solutions rooted in the vast pool of available data. This work presents a comprehensive system for predicting crops and fertilizers based on historical data, leveraging machine learning (ML) algorithms. This work involved research analysis of various ML algorithms to predict crops and recommend suitable fertilizers. It was observed that Naive Bayes and random forest models achieved an excellent accuracy of 99.54% and 99.31%, respectively, for soil classification, indicating their proficiency in distinguishing different soil types. The proposed system also suggests fertilizer recommendations tailored to each crop based on user-provided input and a comparative evaluation of the algorithms. These results highlight the potential of ML techniques in aiding farmers to make informed decisions about soil management and fertilizer selection.