Modeling Daily Stem Water Content by Artificial Neural Network
The purpose of this paper was to model the daily stem water content with neural network. The output voltage of stem water content sensor changed as time series. In order to ensure the accuracy of the model, coefficients sc and eg must be adjusted with the RBF NN input vector changed. The dimensions of input vectors were grouped from 2, 4, S separately. After being grouped, observed data were input to the MATLAB neural network toolbox. To identify all parameters in the model, another set of observed data were used for testing the model. The result shows that the predicted values agreed well with the observed ones. The method of modeling daily stem water with neural network was available. It also indicated that daily stem water content model was not the better with the larger observed data dimensions. Since being measured hourly, the changes of daily stem water content has close correlation with the observed values in 2 hours. It will produce important basis for the precise irrigation system and significantly reduce the logging data storage capacity, greatly improve the computing speed.
artificial neural networks (ANN) stem water content radial basis function neural network (RBF) input vector
Hailan Wang Ye Tian Yandong Zhao
Box 8, Beijing Forestry University Beijing, China
国际会议
成都
英文
414-417
2010-07-07（万方平台首次上网日期，不代表论文的发表时间）