A Novel Hybrid Clustering and Ensemble Learning Framework for Agricultural Crop Recommendation

Selecting a suitable crop for a given soil and climatic condition is an important decision-support task in precision agriculture. This paper investigates a hybrid data-mining framework that combines unsupervised clustering with ensemble classification for crop recommendation. The dataset contains 2,200 records, seven numerical attributes—nitrogen (N), phosphorus (P), potassium (K), temperature, relative humidity, soil pH, and rainfall—and 22 crop labels. After stratified train-test partitioning, K-means clustering was fitted on standardized training features and its cluster assignment was appended as a one-hot encoded feature. A soft-voting ensemble comprising Random Forest, Extra Trees, and Histogram-based Gradient Boosting was then trained to predict the crop class. The proposed cluster-aware ensemble achieved 99.09% accuracy and a weighted F1-score of 99.09% on the 440-record test set. Extra Trees, used as a benchmark, achieved 99.55% accuracy, indicating that clustering augmentation did not outperform every individual model on this dataset. The results demonstrate the feasibility of the hybrid approach but also highlight the need for repeated cross-validation, external validation, and agronomic testing before deployment. The study contributes a reproducible hybrid pipeline and an evaluation protocol rather than claiming that the approach is universally superior.