最优路径森林分类算法综述

Review on optimum-path forest classification algorithm

针对快速分类算法中最优路径森林(OPF)分类算法进行了研究,进行了OPF分类算法研究及应用现状的调查。OPF算法是近期兴起的一种基于完全图的分类算法,在一些公共数据集上与支持向量机(SVM)、人工神经网络(ANN)等算法的对比中,该算法能取得类似或更好的结果,速度更快。该算法不依赖于任何参数、不需要参数优化、不需要对各类别的形状作任何假设,能够处理多类问题,旨在全面系统地介绍OPF算法的研究及应用进展。

This paper did the research on optimal-path forest(OPF)classification algorithm for fast classification algorithm.It investigated the research and application of the OPF classification algorithm.The OPF algorithm is a new classification algorithm based on complete graph.In some public data sets,OPF was compared with support vector machine(SVM)and artificial neural network(ANN),the OPF algorithm could achieve similar or better results,but faster than them.The OPF algorithm does not depend on any parameters,does not need parameter optimization,and also can solve any problems without making any assumptions about the shape of each class.This paper aims to introduce the research status and future research directions of the OPF algorithm to the domestic readers.