LSTM-RNN: An Ensemble Machine Learning Approach for Crop Prediction based on Soil and Weather Properties

This research paper deals with cutting-edge machine learning (ML) techniques to boost agricultural productivity by providing precise crop recommendations tailored to specific soil and weather conditions. The proposed framework merges Long Short-Term Memory (LSTM) networks and Recurrent Neural Networks (RNN) for reliable weather forecasting, while Random Forest (RF) classifiers are employed for the selection of crops. The study validates the effectiveness of this ensemble approach in optimizing crop selection, which significantly outperforms traditional methods. Utilizing a comprehensive dataset spanning 17 years, the model's predictive capabilities are compared with various ML techniques, showcasing its enhanced performance in reducing errors and improving crop yield predictions. This work signifies a notable advancement in precision agriculture, paving the way for intelligent and scalable solutions to address the challenges of global food security.