| 庄哲民,李卡麟,张新蜂,李芬兰.用于早期火灾分类的非线性决策树支持向量机[J].火灾科学,2009,18(4):206-211. |
| 用于早期火灾分类的非线性决策树支持向量机 |
| Nonlinear decision tree support vector machine for early fire classification |
| 投稿时间：2009-08-28 修订日期：2009-09-27 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI：10.3969/j.issn.1004-5309.年.期.顺序 |
| 基金项目:广东省科技计划资助项目(0711050600004) |
|
| 中文关键词: 早期火灾分类 非线性决策树 支持向量机 |
| 英文关键词:Early fire classification Nonlinear decision tree Support vector machine |
| 摘要点击次数: 451 |
| 全文下载次数: 832 |
| 中文摘要: |
| 对早期火灾信息进行研究,提出了一种基于非线性决策树的支持向量机多类分类模型。该模型利用非线性映射将样本投影到高维特征空间,比较每类样本在高维空间的分布情况,进行聚类构造出一个二叉决策树,使容易区分的类别从根节点逐层分类出来,有效克服了错分积累和避免不可分情况;同时,各个节点采用二值最小二乘小波支持向量机,以获得较高的泛化能力。该文将该模型用于早期火灾分类,并与BP神经网络、K近邻法和决策树方法进行比较,实验结果表明,该模型对早期火灾的识别率更高。 |
| 英文摘要: |
| To research into early fire information,a support vector machine multi-class classification model based on nonlinear decision tree is proposed.This model uses nonlinear mapping to map samples into high dimensional feature space,and compares the distribution of each type of samples in the high dimensional space.So that a binary decision tree is constructed through clustering,which can make the more separable classes classified at the upper node of the binary decision tree.The model overcomes the accumulation of misclassification and avoids the condition of impartibility efficiently.At the same time,each node of decision tree uses binary least squares wavelet support vector machine to obtain higher generalization ability.The model is used to classify early fire compared with BP neural network,K-Nearest Neighbor algorithm and decision tree algorithm in this paper.The experiment results show the model has higher recognition rate in the early fire classification. |
| 关闭 |
|
|
|
