Study On Prediction Model Of Physical And Mechanical Wood Properties Of Chinese Fir Plantation Based On Python Language

Chinese fir is an important commodity material with high strength-to-weight ratio and good processability.It has been widely used in home decoration,furniture manufacturing,shipbuilding,bridges,construction and so on.In China,a large area of plantation Chinese fir,was planted in the 1960 s However,the wood of the plantation grows fast,the material is soft,and the mechanical strength is soft.As a consequence,there is a lack of research on the prediction of wood physical and mechanical properties.On the basis of determining the anatomical properties of Chinese fir,the physical and mechanical properties of Chinese fir of plantation are predicted and evaluated by using the relevant algorithms of Python,so as to provide a scientific method for the prediction of wood properties of Chinese fir of plantation.The result shows:(1)The average tracheid lengths of the early wood and late wood of Kai 3,Kai 13,Daba 8,F24 x Na 1-1 are: 3237.18 um,3043.25 um,2664.65 um,2865.82um;3264.05 um,3051.26 um,2709.25 um 2906.34 um.The tracheid length gradually increases from the medulla to the outside.The average value of tracheid length of’Open 3’clones was larger,and the coefficient of variation of tracheid length between early wood and late wood was the smallest.Kai3,Kai13,Daba8,F24 x Na1-1,the average values of microfibril angles of early wood and late wood are: 13.76°,14.62°,15.3°,15.67°;13.04°,14.13°,15.43°,13.54°.The microfibril angle gradually decreases from the medullary core outward.The’open 3’microfibril angle of clones is smaller,and the coefficient of variation of the microfibril angle between early wood and late wood is the smallest.Kai3,Kai 13,Dam 8,F24 x that 1-1 basic density: 0.28 g/cm3,0.31 g/cm3,0.30 g/cm3,0.29 g/cm3.Kai 3,Kai 13,Dam 8,F24 x that 1-1 basic density,gradually increasing from heartwood to sapwood.(2)The model is studied with annual rings,fiber length,microfibril angle,clones and tree species as dependent variables and density as independent variables.The multivariate linear regression algorithm,the neural network regression algorithm of multilayer perceptron(MLP),the decision tree algorithm and the frequently used ensemble algorithm-random forest algorithm are used respectively.Each model is verified many times.As a result,it shows that the average score of multiple linear regression algorithm,multilayer perceptron(MLP)neural network regression algorithm,decision tree and random forest are 0.46,0.61,0.81 and 0.91 respectively.Thus it can be seen that the data is more suitable for the establishment of nonlinear model,and the random forest is more suitable for the establishment of model prediction in the case of a small amount of data.(3)The model is studied with annual rings,fiber length,microfibril angle,clones and tree species as dependent variables and MOR as independent variables.The average score of the decision tree is 0.85,and the average score of the random forest is 0.89 according to the reference of the density model,and using the decision tree and the integrated algorithm random forest.Thus it can be seen that the nonlinear random forest algorithm model is more suitable.(4)The data measured in this paper are added to the Excel table.In addition,the established random forest model is accessed by code.The error between the predicted value of Chinese fir density and the measured value is 9%,and the error between the predicted value of Chinese fir wood MOR and the measured value is 11%.It shows that the random forest algorithm model can be used to accurately predict wood density and MOR on the basis of measuring and analyzing the anatomical properties of plantation Chinese fir.