一种启发式的支持向量机多分类层次树结构构造方法

A heuristic multi-class hierarchical tree structure construction method of support vector machine

支持向量机在处理大样本数据多分类问题时,往往会出现数据集偏斜、基分类器过多、编码方法适应性差等问题,从而导致精度和效率大幅下降。本文提出了一种启发式的支持向量机多分类层次树结构构造方法,构造出由若干子分类器协同计算的多分类层次树结构,一定程度上改善了基分类器过多、数据集偏斜、数据训练时间过长等问题。通过实验证明,该方法与常规方法相比,具有更优秀的精度和效率,在处理大样本、多特征数、多类别的样本分类时,优势更加明显。本研究方法为支持向量机多分类方面的研究提供了一种新的思路,对于本体领域内的分类和回归问题的研究也具有一定的理论参考意义。

For SVMs dealing with multi-classification problems of large sample data,there are often problems such as data set skew,excessive base classifiers,and poor adaptability of coding methods,resulting in a significant drop in accuracy and efficiency.This paper proposes a heuristic method of structure construction for multi-class hierarchical tree via support vector machine.By this method,a multi-class hierarchical tree structure coordinated by several sub-classifiers are constructed, and to some extent it overcomes some problems common in traditional algorithms.The experiment shows that the method has better precision and efficiency than the conventional method,and these advantages are more obvious when dealing with large samples,multi-feature number and multi-class sample classification.This research method provides a new idea to the research of support vector machine multi-classification and manifests certain theoretical significance for the research of classification and regression problems in ontology domain.