Diameter Distribution Prediction of Populus Shelterbelts Based on Artificial Neural Network
Diameter distribution is used to predict stand stock, timber volume and stand yield in most forest management. In the paper, opulus shelterbelts in Boai County were analyzed. A model to predict stand diameter distribution was constructed with artificial neural network(ANN) approach by using the average stand diameter, the coefficient of variation of diameter as well as relative diameter as input variables, and cumulative frequency of tree number as output variables. The structure of the optimum model was 3: 11: 1 and the total fitting accuracy was 98.18 ％. With respect to the model, the max, min and average fitting accuracy of the accumulated frequency could be calculated, which is 99.93 ％, 88.48 ％ and 98.20 ％. The corresponding prediction accuracy was 99.70 ％, 94.36 ％ and 97.56 ％, which has a similar characteristic to that of fitting accuracy. As in forestry practice, a model whose average accuracy is 95％ is reliable enough to meet the practice request. Consequently, it can be concluded that the model works quite well and ANN approach can be used to perform the nonlinear systems such as the complicated stand diameter distribution.
Artificial Neural Network (ANN) Populus Diameter Distribution Prediction Shelterbelts
Fang Guo Guangqin Gao Jiarong Huang Liuxi Wang Dandan Wang
College of Forestry Henan Agricultural University Zhengzhou, 450002, P.R.China
国际会议
2011 Seventh International Conference on Natural Computation(第七届自然计算国际会议 ICNC 2011)
上海
英文
781-784
2011-07-26（万方平台首次上网日期，不代表论文的发表时间）