Soil and Climate Parameters Based Crop Recommendation Model for Yandev, Gboko Local Government Area of Benue State

The increase in human population on fixed landmarks brings about recurrent cultivation of farmlands, which degrades the soil. The varied nature of the soil and climate escalates this situation. Making farming decisions based on human experience is misleading, especially in the everchanging climates. Farmers now experience low agricultural output in the Yandev District of Benue State, Nigeria. Hence, proper management of farming resources with the aid of artificial intelligence has become pertinent. Existing solutions may not be effective in the Yandev axis due to the geographic disparity of the datasets used to train the models. This study leveraged soil and climate parameters in the Yandev district to develop a crop recommendation model. Five machine learning models, including decision trees, naive Bayes, logistic regression, random forest, and extreme gradient boosting, were utilized. The models were experimented with the Yandev dataset and evaluated using accuracy,
$\mathbf{F}1$
-score, precision and recall metrics. The evaluation shows that the naive Bayes algorithm outclassed the other learning models with a performance of
$99.85\%$
accuracy. This experimental result also outperformed other existing methods, indicating the significance of the model in the area. The goal of recommending what crop to grow in the Yandev axis was achieved.