Chance constrained conic-segmentation support vector machine with uncertain data

TL;DR

This paper extends the conic-segmentation SVM to handle uncertain or mislabelled data using chance constraints, providing a probabilistic guarantee of low misclassification rates and demonstrating its effectiveness through experiments.

Contribution

It introduces a chance-constrained approach to conic-segmentation SVM for uncertain data, enhancing robustness and reliability in classification.

Findings

Chance-constrained CS-SVM reduces misclassification probability.

The geometric interpretation clarifies the model's operation.

Experimental results show improved performance with uncertain data.

Abstract

Support vector machines (SVM) is one of the well known supervised classes of learning algorithms. Furthermore, the conic-segmentation SVM (CS-SVM) is a natural multiclass analogue of the standard binary SVM, as CS-SVM models are dealing with the situation where the exact values of the data points are known. This paper studies CS-SVM when the data points are uncertain or mislabelled. With some properties known for the distributions, a chance-constrained CS-SVM approach is used to ensure the small probability of misclassification for the uncertain data. The geometric interpretation is presented to show how CS-SVM works. Finally, we present experimental results to investigate the chance constrained CS-SVM's performance.