摘要
应用ARIMA模型（差分自回归移动平均模型）对圈养亚成体大熊猫24h的取食行为节律时间序列进行分析和预测。根据序列图判断时序数据的平稳性,结合自相关及偏自相关函数确定分析模型为ARIMA（p,d,q）（P,D,Q）~8模式。基于Expert Modeler模块和AIC信息评判原则筛选ARIMA模型的最佳阶数组合为ARIMA（0,0,1）（1,1,0）~8。模型参数估计均通过显著性检验,预测值的置信区间包括所有实际观测数值。预测值残差的自相关与偏自相关函数值均在置信区间内,残差为白噪声,模型的检验结果比较理想和有效。
We used the ARIMA model（Autoregressive Integrated Moving Average Model）to analyze and predict the feeding behavior of captive sub-adult giant panda during a 24-hour period.We determined the stability of time series data using sequence charts,and the ARIMA（p,d,q）（P,D,Q）＇ model was selected based on the autocorrelation function（ACF） and partial autocorrelation function（PACF）.Through Expert Modeler and AIC information criterion,the model was set as ARIMA（0,0,1）（1,1,0）~8.The modeled parameter estimate was supported by significance tests,and the observed data fell within the confidence interval predicted by the model.According to the ACF and PACF,the residuals were randomly distributed, and the calculation results of the ARIMA model were correct and effective.
出处
《野生动物》
2011年第4期183-185,220,共4页
