基于K-近邻的局部懒惰式决策树分类模型

A Local Lazy Decision-tree Classifier Based on K-nearest Strategy

分类是数据挖掘的一个重要研究课题,其概念是在已有数据的基础上构造出一个分类模型。该模型能够把数据库中的数据记录映射到给定类别中的某一个,从而进行数据的分类。通过对懒惰式学习策略的研究,在大量实验的基础之上,提出了一个新的分类模型——Local-LDtree。介绍了Local-LDtree模型的原理和算法,分析了其在分类精确度方面的优劣,指出了对其进行改进的方向。

Classification is one of very important basic tasks in the field of data mining,and a task of classification is to find a classification model.This model can map an unlabelled instance,which is represented by a set of attribute-value pairs,into its predicted label.Such a model can be applied to data classification.On the basis of analyzing lazy learning strategy and on the results of the experiments,this paper proposes a new lazy classification model,Local-LDtree,and introduces the principle and algorithm of Local-LDtree,analyzes the classification accuracy of the model,points out the direction about how to improve the model.