An Intelligent Crop Recommendation Framework Using a Deep Learning and Metaheuristic Optimization Approach

Crop recommendation systems examine variables like soil type, climate, and historical data to recommend appropriate crops for a given area with the aid of various machine learning and deep learning models. This study suggests a novel crop recommendation system that makes use of a Deep Convolutional Neural Network (DCNN). The Modified Salp Swarm Algorithm (MSSA), which depends on a competition mechanism and variable shifted windows, is used to optimize the DCNN's weight update. By adding a competitive mechanism that dynamically ranks salps depending on their fitness and permits stronger people to advise weaker ones, the MSSA improves the traditional Salp Swarm Algorithm and prevents early convergence. Concurrently, the variable shifting window technique improves exploration in the early stages and exploitation near convergence by adaptively modifying the search bounds of salps during iterations. Across several benchmark datasets, results from experiments verify that DCNN–MSSA performs noticeably better than traditional optimizers when measured in terms of performance metrics.