摘要： 分类是数据挖掘、机器学习和模式识别中一个重要的研究领域。通过对当前数据挖掘中具有代表性的优秀分类算法进行分析和比较,总结出了各种算法的特性,为使用者选择算法或研究者改进算法提供了依据。此外,提出了评价分类器的5条标准,以便于研究者提出新的有效算法。

关键词: 数据挖掘, 分类, 算法, 数据集

Abstract: Classification is an important research field in data mining, pattern recognition and machine learning. This paper summarizes the main features of every algorithm by analyzing and comparing a variety of typical classifiers to provide a basis for selecting or improving the algorithms in data mining. Besides, it puts forward five standards according to evaluating classifiers so that researchers develop new effective algorithms.

中图分类号: 

• TP311.13