Application of Kernel Density Estimation Theory to Establishment of Tree Heightdiameter Curve

On the basis of analyzing the disadvantage of the Least Squares (LS) fitting method, the heightdiameter curve model for Pinus keysia var. langbianensis tree species in Yunnan Province was established by means of nonparametric Kernel Density Estimation method. It was showed by the study that the result from the Kernel Density Estimation method was better than LS method when the appropriate kernel function and band width were chosen. Kernel Density Estimation method can be applied not only to model establishment, but also to data test. It was showed that the data sample should be large enough for model building with Kernel Density Estimation method, otherwise the data fluctuation of the model would be beyond expectation.