Research On Growth Model Of Korean Pine Based On Combined Model

This paper mainly studies the tree height growth model of natural Korean pine forest in Liangshui Experimental Forest Farm in Yichun City,Heilongjiang Province.The model uses the analytical data of 4 trees in 4 plots in the forest farm.Based on traditional tree growth theory,time series theory,neural network theory and combination theory,an optimal weighted combination model describing the growth law of tree age and tree height is given.The analysis proves that this model is the best model to describe the natural Korean pine tree age-height growth law.First,establish a longitudinal growth model.For the high growth data of different red pine trees,2 sets of data are reserved for each plant for verification,and the remaining data are used for fitting.Establish a traditional tree growth model,and based on this,combine the autoregressive model into a residual autoregressive model.The residual autoregressive model has the interpretability of the parameters of the traditional tree growth model and the autoregressive model has the characteristics of removing the autocorrelation of the data.Then establish the BP neural network model,and use the genetic algorithm(GA)to optimize its initial weights and thresholds to obtain the GA-BP neural network model.GA-BP neural network model has the nonlinear fitting ability of BP neural network and the global optimization ability of genetic algorithm.The optimal weighting method is used to combine the residual autoregressive model and the GA-BP neural network model to obtain the optimal weighted combination model.After comparison and analysis,the results show that the fit degree and verification accuracy of the optimal weighted combination model is closer to the height growth law of red pine than other models,and it is the optimal longitudinal growth model.Secondly,establish a lateral growth model.The height and growth data of the first three trees were used for fitting,and the data of the fourth tree was used for verification.The traditional theoretical growth model,autoregressive model and GA-BP neural network model are established respectively,and then the three models are combined to obtain the weight by using the optimal weighting method to obtain the horizontal combined growth model.The results show that the fit degree and verification accuracy of the optimal weighted combination model are closer to the height growth of Korean pine than other models,and it is the optimal lateral growth model.Finally,through comparative analysis,it is confirmed that the optimal weighted combination model is the optimal model describing the age-height growth law of Korean pine.