ywh19860616 发表于 2011-12-31 10:27
epoh老师,还有一个问题,您有空帮忙看看,谢谢了
原程序只是滞后一阶,而我现在想修改滞后阶数,出现了很 ...
select ttrial=ceil(pc*ntrial/100),
此处100跟抽样无关
假设ntrial=100,而pc=1 TO 100;
当pc=95,pc*ntrial/100=95,第95个就是wald95
当pc=99,pc*ntrial/100=99,第99个就是wald99
同理ntrial=10000,而pc=1 TO 100;
当pc=95,pc*ntrial/100=9500,第9500个就是wald95
当pc=99,pc*ntrial/100=9900,第9900个就是wald99
假设你不想把cdfwald存成文件,
也可以改成底下:
MAT tcdfwald=cdfwald';
print tcdfwald;
unmake tcdfwald wald1-wald100;
print wald90;
print wald95;
print wald99;
#######
滞后阶数修改,这两天我试试.
#######
freq(panel,n=29,t=38,id=country,time=year) n;
set nt = 29*38;
smpl 1,nt;
read(file='c:\wald1.xls');
print @NOB;
@NOB = 1102.00000
######
TREND generates a series with a linear growth trend.
The trend may be repeated under the control of several options.
ywhtsp.tsp
FREQ A;
SMPL fyear lyear;
TREND t;
T
1960 1.00000
1961 2.00000
1962 3.00000
1963 4.00000
1964 5.00000
1965 6.00000
1966 7.00000
1967 8.00000
1968 9.00000
1969 10.00000
1970 11.00000
1971 12.00000
1972 13.00000
1973 14.00000
1974 15.00000
1975 16.00000
1976 17.00000
1977 18.00000
1978 19.00000
1979 20.00000
1980 21.00000
1981 22.00000
1982 23.00000
1983 24.00000
1984 25.00000
1985 26.00000
1986 27.00000
1987 28.00000
1988 29.00000
1989 30.00000
1990 31.00000
1991 32.00000
1992 33.00000
1993 34.00000
1994 35.00000
1995 36.00000
1996 37.00000
1997 38.00000