Neo-Twin Support Vector Machines for Pattern Classification
Sambhav Jain, Shuvo Saha Roy, Reshma Rastogi · 2022 International Conference on Decision Aid Sciences and Applications (DASA) · 2022
In this paper, a new variant of Twin Support Vector Machines (TSVM) termed as Neo-Twin Support Vector Machines (Neo-TSVM) has been proposed for binary pattern classification. TSVM uses hinge loss to allow optimal separation from the opposite class, casting it as a constrained optimisation problem. Neo-TSVM presents a simpler model which eliminates the constraints and cast it as an Unconstrained Minimisation Problem (UMP). Further to allow, better separation between the non-parallel hyperplanes, the notion of angle has also been introduced in the optimisation problem. For testing the efficacy of the proposed classifier, experiments have been conducted on benchmark datasets, and it is observed that the proposed classifier achieves results comparable to that of TSVM, and is also time efficient.