摘要
分析了木材材质预测研究中常用的三种建模方法:回归分析方法、时间序列方法和神经网络方法的特点,以木材生长轮密度预测为例,给出了三种预测方法的建模过程与结果,并进一步对多个材性指标建模比较分析,认为:神经网络方法是木材材质预测的最优选择。
The characteristics of regression analysis method, time series method and neural network method commonly used in wood quality forecast were analyzed. The modeling process and result of these forecasting methods were presented in terms of the density forecast of wood growth ring. The modeling result is analysed by multi-indexes of wood quality. It indicates that neural network method is the best method for wood quality forecast.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2008年第6期30-31,47,共3页
Journal of Northeast Forestry University
基金
国家自然科学基金(30671645)
黑龙江省科技攻关项目(GC05B605,GC05A118)
黑龙江省自然科学基金(F200506)
关键词
材质预测
神经网络
时间序列分析
回归分析
Wood quality forecast
Neural network
Time series analysis
Regression analysis