Fuzzy least squares projection twin support vector machines for class imbalance learning

In this paper, we propose a novel fuzzy least squares projection twin support vector machines for class imbalance learning (FLSPTSVM-CIL). Unlike machine (TSVM) which solves two dual problems, solve modified primal formulations by solving systems of linear equations. The proposed FLSPTSVM-CIL model seeks directions such that the samples classes are well separated in projected space. To avoid singularity issues, incorporate an extra regularization term to make optimization problem positive definite. As real world data may be imbalanced, assign appropriate weights classifier is not biased towards majority class. statistical analysis and experimental results on publicly available UCI benchmark datasets show performs better as compared baseline models. applications datasets, performed classification Alzheimer’s disease breast cancer patients. Experimental generalization performance patients mild cognitive impairment versus subjects