SoWhat: Real-Time Crop and Fertilizer Predictor

Real-time, automatic crop and fertilizer prediction has the potential to improve crop yield, optimize resource allocation, and enhance farming practices. The proposed work introduces the ‘SoWhat’ mobile application based on an automatic crop and fertilizer predictor that provides accurate real-time crop and fertilizer prediction. It is user-friendly and built using the Flutter interface and a Python script based on the decision forest-based algorithm. Various evaluation metrics, test set evaluation, and a comparison examination of seven machine learning algorithms have all been used to demonstrate the application’s efficacy. Seven machine learning models—lasso regression, logistic regression, ridge regression, K-neighbors classifier, random forest classifier, decision tree classifier, and linear regression model—were thoroughly compared before the decision forest approach was chosen. The decision tree model’s accuracy, precision, and F1-score were 100.00, 0.98, and 0.98, respectively. In conclusion, this study emphasizes how data-driven approaches and machine learning may revolutionize agriculture, focusing on farmer’s needs and challenges.