Intuitionistic Fuzzy Laplacian Twin Support Vector Machine for Semi-supervised Classification

Abstract In general, data contain noises which come from faulty instruments, flawed measurements or communication. Learning with in the context of classification regression is inevitably affected by data. order to remove greatly reduce impact noises, we introduce ideas fuzzy membership functions and Laplacian twin support vector machine (Lap-TSVM). A formulation linear intuitionistic (IFLap-TSVM) presented. Moreover, extend IFLap-TSVM nonlinear case kernel function. The proposed resolves negative outliers using a more accurate reasonable classifier geometric distribution information labeled unlabeled based on manifold regularization. Experiments constructed artificial datasets, several UCI benchmark datasets MNIST dataset show that has better accuracy than other state-of-the-art (TSVM), (IFTSVM) Lap-TSVM.