SMOTE-based method for balanced spectral nondestructive detection of moldy apple core

Moldy apple core is an internal fruit disease that poses a threat to consumer health. In this study, synthetic minority over-sampling technology (SMOTE) based model of moldy was proposed solve the problem poor performance due imbalanced sample distribution in spectral nondestructive detection core. Two different methods, random under-sampling (RUS) and SMOTE, were used balance original samples. A hybrid strategy competitive adaptive reweighted sampling (CARS) successive projections algorithm (SPA) for variable selection, partial least squares-discriminant analysis (PLS-DA) classification squares-support vector machine (LS-SVM) nonlinear constructed. The SMOTE-based method better than RUS solving distribution. constructed SMOTE-CARS-SPA-LS-SVM could not only accurately identify healthy diseased apples with degree >10% but also improve accuracy slightly addition, has certain practical applicability external verification.