太阳黑子与杉木生长关系

根据多层误差板传网络结构模型和三次设计发展了一种太阳黑子人工神经网络预报方法，以杉木胸径生长的年轮指数和太阳黑子自相关因子输入变量，应用改进的人工神经网络方法建立了太阳黑子相对数年平均值的预测模型，模型的模拟回归优度为93．3％，预测精度达到95．74％，并对网络模型中变量进行灵敏度分析，分析表明，杉木胸径生长的年轮指数对太阳黑子对相对数年平均值影响较平坦，而太阳黑子自相关因子Yt-4和Yt-2对太阳子相对数年平均值影响较灵敏，3个因子对太阳黑子相对数年平均值均在一定的影响。图2表5参19

THE RELATIONSHIP BETWEEN SUN-SPOT AND GROWTH OF CUNNINGHAMIA LANCEOLATA

A new forecast method of sun-spot is developed based on the artificial neural network of the back propagation model. In this paper, the modified artificial neural network method is used to establish forecast model of sun-spot, regarding the ring index of diameter at breast height of Cunninghamia lanceolata and auto-correlated factors of sun-spot as forecast factors. The simulating regression precision of model is 93.3%, and the predicting precision of model is 95.74%. Based on the artificial neural network model, the sensitivity analysis on each variable of model shows that the effect of ring index of diameter at breast height of Cunninghamia lanceolata on sun-spot is even, and the effects of auto-correlated factors of sun-spot on sun-spot are sensitive, which shows that three variables all have significant effects on sun-spot. Fig 2, Tab 5, Ref 19