请教一下高手，我做时间序列分析，经过一阶差分以后，确定了ARIMA模型为（1，1，1）。但是我在做estimate这个步骤时，报错

ERROR:

The estimation algorithm did not converge after 50 iterations.

请教高手

proc arima
data=shouru.a;

identify var=x(1) nlag=12;

run;

estimate
p=1 q=1;

run;

forecast lead=12 interval=month id=date out=results;

run;
我这个程序有什么问题啊？谢谢啦.为何我运行出来不预报啊。
出来的是这个
ARIMA Estimation Optimization SummaryEstimation MethodConditional Least SquaresParameters Estimated3Termination CriteriaMaximum Relative Change in EstimatesIteration Stopping Value0.001Criteria Value0.047823Alternate CriteriaRelative Change in Objective FunctionAlternate Criteria Value0.000518Maximum Absolute Value of Gradient1.2294E9R-Square Change from Last Iteration0.129779Objective FunctionSum of Squared ResidualsObjective Function Value2.6653E9Marquardt's Lambda Coefficient1E-12Numerical Derivative Perturbation Delta0.001Iterations50Warning MessageThe estimation algorithm did not converge
如果我是运行

proc arima
data=shouru.a;

identify var=x  nlag=12;

run;

estimate
p=1 q=1;

run;

forecast lead=12 interval=month id=date out=results;

run;
就会出来没有经过1阶差分的，p=1，q=1的预报结果

而且这个程序，也会出来预报结果;
proc arima data=tax.yijie;
identify var=x(1,12) nlag=12 ;
estimate p=1 q=1 method=cls;
forecast lead=12 interval=month id=date out=results;
run;


这是为什么呢？
希望各位帮帮忙啊