Crop Connect: An Approach to Crop Recommendation

This paper presents a crop recommendation system that incorporates five essential factors namely NPK values, temperature, humidity, pH, and rainfall. The purpose of this research is to predict the appropriate crop based on these conditions. In order to achieve this goal, machine learning algorithms, GradientBoostingRegressor and XGBoost have been used for rainfall prediction and crop recommendation respectively. The system's methodology involves collecting data on NPK value, temperature, humidity, pH, and historical rainfall patterns. Using the GradientBoostingRegressor model, rainfall value is accurately predicted which serves as a crucial input for the subsequent crop recommendation process. By employing XGBoost, the combined influence of the NPK value, temperature, humidity, pH, and predicted rainfall is analyzed to determine the most appropriate crop for a given set of conditions. The proposed crop recommendation system offers a practical solution for farmers and agricultural professionals to optimize crop selection and maximize yields. It considers multiple factors simultaneously, providing unique insights and enhanced decision-making in crop planning. The average root mean square error measured over 4 months for the model used in the system is 0.19 which indicates a high level of accuracy and precision in the prediction of rainfall, thus producing reliable crop recommendations. This paper's findings and methodologies contribute to the field of precision agriculture and provide a valuable tool for sustainable and efficient crop cultivation.