FITTING Data Mining Settings for Ranking Seed Lots

To enhance speed and agility in interpreting physiological quality tests of seeds, The use algorithms has emerged. This study aimed to identify suitable machine learning models assist the precise management seed lot quality. Soybean lots from two companies were assessed using Supplied Test Set, Cross-Validation (with 8, 10, 12 folds), Percentage Split 66% 70%) methods. Variables analyzed through Tetrazolium included vigor, viability, mechanical damage, moisture bed bug water content. Method performance was determined by Kappa, Precision, ROC Area metrics. Classification Via Regression J48 employed. technique utilizing data for training achieved 93.55% accuracy, with Precision reaching 94.50% algorithm. Applying cross-validation method 10 folds resulted 90.22% correctly classified instances, a outcome like previous method. Vigor primary attribute used. However, these results are specific this study's database, careful planning is necessary select most effective application