Research And Forestry Application On Non-parametric Local Polynomial Model

Forest ecosystem management and providing ecological services has become hot spots of forest ecosystems, resulting in forest ecosystem management model, growth curve models are mostly used, so the study of the growth curve model, the establishment of a reasonable growth curve model is the focus of the forest management. The growth curve model is mostly parametric model, as the parameters can effectively explain the biological significance of the model. This paper discusses the simulation of growth curves using non-parametric methods, compared to the conventional parameters, provides another idea of simulating the growth curve data.In this paper, we study on the basic principle of the non-parametric local polynomial regression model, with reference to the SAS software nonparametric local polynomial regression model, to build the basic framework of the non-parametric local polynomial regression model and algorithm steps, to complete the weights choice of model, the optimal choice of bandwidth and the judgement of MSE based on GCV, to complete the non-parametric local polynomial regression model. Instance verifies the model results are same as the SAS simulation. In this paper, the realization of the local polynomial model is based on forestry Statistics software Forstat software platform by using the underlying matrix operations function. In addition, the growth curve data can be automatically generated by this model for forest management software FSOS, which makes the FSOS growth curve data entry more convenient.