Uma heurística probabilística para o problema de Classificação Geodésica

Abstract

This work proposes a probabilistic heuristic for the 2-Class Single Group Geodesic Classification Problem (2-SGGC), which was introduced in Araújo et al. (2019a), where the authors present an integer linear programming model to solve it exactly. The problem aims to classify data by recognizing patterns that were defined by geodesic convexity from pre-established information and represented by a binary relationship using a graph. We performed computational experiments to evaluate the computational efficiency and classification accuracy of the proposed approach, comparing it with some classic methods of solution for the classification problem. It is worth mentioning that our heuristic is an improvement of the heuristic developed for the 2-Class Multi Group Geodesic Classification problem in Vieira (2019).