Recognition of the hardness of licorice seeds using a semi-supervised learning method and near-infrared spectral data

► We propose a new semi-supervised support vector machine ( S3VM) formulation. ► We identify the hardness of licorice seeds employing S3VM and NIR spectral data. ► DC programming is used to solve the resulting model. ► The proposed method converges quickly and has low computational burden. ► Experimental results show that our S3VM outperforms the existing methods.