【作者】 黄家荣； 郭芳； 王留喜；

【Author】 Huang jia-rong Guo fang Wang liu-xi

【机构】 河南农业大学林学院；

【摘要】 以河南省博爱县毛白杨农田防护林带为研究对象,在调查、计算林带径阶分布数据和有关林分特征数的基础上,应用人工神经网络建模技术,以断面积平均直径、直径变动系数和径阶相对直径等为输入变量,以径阶株数累积频率为输出变量,建立了结构为3:11:1的毛白杨农田防护林带直径分布神经网络预测模型,模型总体拟合准确度为98.18%。模型的拟合准确度分析结果表明,累积频率拟合准确度最大99.93%,最小88.48%,平均98.20%。模型的预测准确度分析结果显示,累积频率预测准确度最大99.70%,最小94.36%,平均97.56%。模型的平均准确度在95%以上,能满足林业上的应用要求,模型研究取得了较好的拟合、预测效果。更多还原

【Abstract】 A model forecasting stand diameter distribution was created by using artificial neural network modeling technology, with an average diameter and the coefficient of variation in diameter and diameter opposite order for input variables in diameter, number of cumulative frequency for advanced output variable, in Populus shelterbelts in Bo Ai. County.Through training and optimum seeking, the idea model was created, in which the model structure is 3:11:1, the total fitting accuracy is 98.18 %. Concretely, the accumulated frequency fitting accuracy is 88.48 % to 99.93 % and the mean value is 98.20 %. The cumulated frequency prediction accuracy is 94.36 % to 99.70 % and the mean value is 97.56 %.The mean accuracy of the model is over 95% so that it can be used in forestry and the effect is better.更多还原