摘要
用BP人工神经网络和Richards方程2种方法建立了马尾松人工林地位指数模型,其中Richards方程的参数由改进的单纯形法求解。研究结果表明,用人工神经网络建立模型拟合精度为94.52%,检验精度为92.15%,所建模型无系统偏差,与Richards方程相比,具有较好的拟合效果,人工神经网络更适合建立非线性模型。
The paper modeled the site index of masson pine fortunei plantation with BP Neural Network and Richards formula whose parameters were calculated by improved simplex method. The results demonstrated that the fitting precision of BP Neural Network model was 94.52% and the test precision was 92.15%, the established model possessed better drawing up effect than Richards formula, so BP Neural Network was more suitable to establish nonlined model.
出处
《西北林学院学报》
CSCD
北大核心
2009年第2期128-130,共3页
Journal of Northwest Forestry University
基金
河南省科技攻关基金课题(0624050007)
