10th U6+ Consortium of African Universities International Conference

Crop recommendation system an innovation in precision agriculture, plays a critical role in improving productivity by identifying suitable crops based on soil and climatic conditions. This study presents a comparative analysis of supervised machine learning algorithms for innovative crop recommendation using agro-environmental parameters, including Nitrogen, Phosphorus, Potassium, temperature, humidity, rainfall, and soil pH. The dataset used is the publicly available crop recommendation dataset obtained from Kaggle, and was filtered to include crops commonly grown in Nigeria, resulting in approximately 900 observations. Five supervised machine learning algorithms including Logistic Regression, Decision Tree, Random Forest, Bagging Classifier, and Support Vector Classifier, were evaluated using a 10-fold cross-validation approach to ensure reliable performance assessment. The results showed that all models achieved high classification accuracy, with Decision Tree, Random Forest, and Bagging Classifier attaining perfect accuracy, while Logistic Regression and Support Vector Classifier achieved slightly lower accuracy scores. The high performance observed across all models is attributed to the structured nature of the dataset and the strong relationship between soil and climatic features and crop suitability. However, the near-perfect accuracy highlights the need for cautious interpretation, as it may reflect dataset characteristics rather than real-world complexity. Overall, the study demonstrates the effectiveness of supervised machine learning algorithms for crop recommendation system and identifies ensemble methods as the most robust and consistent performers.