Abstract

Soil quality is an important aspect in successful agriculture, impacting crop output and sustainability. Traditional soil prediction methods frequently encounter uncertainty and inadequate information. This study provides a unique way to improving soil quality prediction by combining fuzzy soft sets with machine learning algorithms. Using a complete dataset encompassing variables such as soil pH, nutrient levels (Nitrogen, Phosphorus, Potassium), temperature, humidity, and rainfall, our solution overcomes the inherent uncertainties in soil data by representing imprecise and ambiguous properties with fuzzy logic. Soft sets are used to accommodate partial data, giving a solid foundation for decision-making. Machine learning models, such as regression and classification approaches, are then trained on the improved information to forecast soil quality outcomes.