home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Frostbyte's 1980s DOS Shareware Collection
/
floppyshareware.zip
/
floppyshareware
/
DOOG
/
PCSSP2.ZIP
/
STATNONP.ZIP
/
SMIRN.FOR
< prev
next >
Wrap
Text File
|
1985-11-29
|
3KB
|
86 lines
C
C ..................................................................
C
C SUBROUTINE SMIRN
C
C PURPOSE
C COMPUTES VALUES OF THE LIMITING DISTRIBUTION FUNCTION FOR
C THE KOLMOGOROV-SMIRNOV STATISTIC.
C
C USAGE
C CALL SMIRN(X,Y)
C
C DESCRIPTION OF PARAMETERS
C X - THE ARGUMENT OF THE SMIRN FUNCTION
C Y - THE RESULTANT SMIRN FUNCTION VALUE
C
C REMARKS
C Y IS SET TO ZERO IF X IS NOT GREATER THAN 0.27, AND IS SET
C TO ONE IF X IS NOT LESS THAN 3.1. ACCURACY TESTS WERE MADE
C REFERRING TO THE TABLE GIVEN IN THE REFERENCE BELOW.
C TWO ARGUMENTS, X= 0.62, AND X = 1.87 GAVE RESULTS WHICH
C DIFFER FROM THE SMIRNOV TABLES BY 2.9 AND 1.9 IN THE 5TH
C DECIMAL PLACE. ALL OTHER RESULTS SHOWED SMALLER ERRORS,
C AND ERROR SPECIFICATIONS ARE GIVEN IN THE ACCURACY TABLES
C IN THIS MANUAL. IN DOUBLE PRECISION MODE, THESE SAME
C ARGUMENTS RESULTED IN DIFFERENCES FROM TABLED VALUES BY 3
C AND 2 IN THE 5TH DECIMAL PLACE. IT IS NOTED IN
C LINDGREN (REFERENCE BELOW) THAT FOR HIGH SIGNIFICANCE LEVELS
C (SAY, .01 AND .05) ASYMPTOTIC FORMULAS GIVE VALUES WHICH ARE
C TOO HIGH ( BY 1.5 PER CENT WHEN N = 80). THAT IS, AT HIGH
C SIGNIFICANCE LEVELS, THE HYPOTHESIS OF NO DIFFERENCE WILL BE
C REJECTED TOO SELDOM USING ASYMPTOTIC FORMULAS.
C
C SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED
C NONE
C
C METHOD
C THE METHOD IS DESCRIBED BY W. FELLER-ON THE KOLMOGOROV-
C SMIRNOV LIMIT THEOREMS FOR EMPIRICAL DISTRIBUTIONS- ANNALS
C OF MATH. STAT., 19, 1948, 177-189, BY N. SMIRNOV--TABLE
C FOR ESTIMATING THE GOODNESS OF FIT OF EMPIRICAL
C DISTRIBUTIONS- ANNALS OF MATH. STAT., 19, 1948, 279-281,
C AND GIVEN IN LINDGREN, STATISTICAL THEORY, THE MACMILLAN
C COMPANY, N. Y., 1962.
C
C ..................................................................
C
SUBROUTINE SMIRN(X,Y)
C DOUBLE PRECISION X,Q1,Q2,Q4,Q8,Y
C
C IF A DOUBLE PRECISION VERSION OF THIS ROUTINE IS DESIRED, THE C
C IN COLUMN ONE OF THE DOUBLE PRECISION CARD ABOVE SHOULD BE
C REMOVED, AND THE C IN COLUMN ONE OF THE STATEMENTS NUMBERED
C C 3, C 5, AND C 8 SHOULD BE REMOVED, AND THESE CARDS
C SHOULD REPLACE THE STATEMENTS NUMBERED 3, 5, AND 8,
C RESPECTIVELY. ALL ROUTINES CALLING THIS ROUTINE MUST ALSO
C PROVIDE DOUBLE PRECISION ARGUMENTS TO THIS ROUTINE.
C
C ..................................................................
C
IF(X-.27)1,1,2
1 Y=0.0
GO TO 9
2 IF(X-1.0)3,6,6
3 Q1=EXP(-1.233701/X**2)
C 3 Q1=DEXP(-1.233700550136170/X**2)
Q2=Q1*Q1
Q4=Q2*Q2
Q8=Q4*Q4
IF(Q8-1.0E-25)4,5,5
4 Q8=0.0
5 Y=(2.506628/X)*Q1*(1.0+Q8*(1.0+Q8*Q8))
C 5 Y=(2.506628274631001/X)*Q1*(1.0D0+Q8*(1.0D0+Q8*Q8))
GO TO 9
6 IF(X-3.1)8,7,7
7 Y=1.0
GO TO 9
8 Q1=EXP(-2.0*X*X)
C 8 Q1=DEXP(-2.0D0*X*X)
Q2=Q1*Q1
Q4=Q2*Q2
Q8=Q4*Q4
Y=1.0-2.0*(Q1-Q4+Q8*(Q1-Q8))
9 RETURN
END