应用SOM网络模型进行防火树种分类研究  

A Study on the Application of Network Model of SOM to Classification of Fireproof Trees

作　　者：郑德祥[1] 陈平留[1] 陈文琴[1] 

机构地区：[1]福建农林大学林学院,福建福州350002

出　　处：《江西农业大学学报》2006年第4期554-558,共5页Acta Agriculturae Universitatis Jiangxiensis

基　　金：福建省教育厅科学基金项目(JB200269)

摘　　要：应用SOM神经网络模型,以福建省37个针阔树种的叶子的含水率、粗脂肪、灰分、S iO2、挥发油含量、燃点、燃烧热及燃烧速度为分类因子,进行网络学习与仿真,结合过去的分类方法将37个树种分为5类,从而验证SOM网络模型在防火树种选择分类中的应用效果,以提供一种新的分类思路与方法。By means of nerve network model of SOM （Self- Organizing Feature Map） , 37 species of coniferous and broad - leaf trees in Fujian were divided into five kinds according to the rate of water, the content of crude fat, ashes, SiO2, volatile oil in the leaves,and their ignition point,heat of combustion and speed of burning. The network study and simulation were carried out, in combination with conventional methods, then the effect of SOM on choosing and classifying the fireproof trees was verified, thus providing a new conception and method for classification.

关 键 词：抗火性能 防火树种 SOM模型 模式分类 

分 类 号：F290[经济管理—国民经济]