A hybrid agro-economic framework with elasticity-based demand modeling for intelligent crop recommendation

Crop planning increasingly requires environmental and economic information to support sustainable and profitable agricultural decisions. Existing crop recommendation systems primarily emphasize agroclimatic conditions while overlooking dynamic market factors that influence farmers’ choices. This study presents a hybrid agroeconomic framework integrating weather, soil, and market parameters for intelligent crop recommendation. The dataset combines meteorological variables (temperature, rainfall, humidity), soil data (N, P, K, pH, moisture) and agricultural market indicators (modal price, commodity arrivals) attained from Agmarknet. To address the absence of explicit demand data, an elasticity-based demand estimation approach is introduced to estimate market demand from the changes in supply and price. The proposed method is a novel approach of that combines heterogenous datasets through spatial and temporal alignment into a single feature space. The integrated dataset is then used to train and evaluate different machine learning models. A hybrid ensemble combining Random Forest and Extreme Gradient Boosting (XGBoost) achieved the highest validation accuracy of 95%, outperforming Support Vector Machine, Logistic Regression, and Decision Tree models. The proposed framework provides a scalable and realistic platform for developing cloud-based agricultural decision-support systems, allowing crop recommendations that are both agronomically suitable and economically viable.