estimate the number of source using AIC and MDL criterion.

时间序列分析中AR模型的建立，可求的残差平方和，便于用AIC准则定阶。-Time series analysis of AR model, rectifiable the residual sum of squares, easy to use the AIC criterion to determine the order.

生成AR(p)过程模拟的参数。满足过程的弱稳定。模拟多条样本路径，利用AIC/BIC/IC进行定阶。比较三种标准的准确性。-Generate an AR (p) process simulation parameters. Meet process weakly stable. Simulation sample path, AIC/BIC/IC order determination. Three standard accuracy.

FPE，AIC，CAT三种AR模型预测准则，可对同一个序列同时进行阶数分析并作图，相互比较。-FPE，AIC，CAT order prediction ,can be used on the same time series and contrast with each other