RBF模型在林地利用区域差异判定中的应用

Application of radial basis function networks to regional differentiation of forestland use

应用文献资料法、RBF神经网络法和聚类分析法,以反映林地利用程度的经济指标为样本对福建省68个县市的林地利用区域差异进行评价研究。结果表明,福建省林地利用区域差异明显,对RBF模型评价结果进行聚类分析可将福建省林地利用水平分为3级,林地利用水平较高的县市主要分布于闽北与闽西北,林地利用水平较低的县市主要分布于闽东及闽南沿海县市,而闽西与闽中地区林地利用水平介于二者之间,其评价结果与福建省林地实际利用状况基本相符。由此可见,RBF模型可应用于评价林地区域差异与利用水平,具有良好的发展和应用前景。

The purpose of this study is to set up a radial basis function(RBF) neutral network model by economic indices of forestland use degree for evaluating the forestland use level and regional differentiation in 68 counties or cities of Fujian Province,southeastern China.Methods of documentation,RBF neutral network and cluster analysis were employed.The results indicate that the regional differentiations of forestland use in Fujian are marked.The clustering method was adopted in the evaluation value of RBF model and the forestland use of Fujian Province was classified into three groups.Counties or cities with higher evaluation values were mainly distributed in the northern and northwestern parts,and those with lower values primarily in the eastern and the southern coastal areas,but the forestland use level in the western and the middle areas of Fujian was medium.Basically,the evaluation value accords with the real situation of the forestland use in Fujian,which also indicates that application of RBF model to assessing regional differentiations and the use level of forestland is feasible and can be well developed and promoted.