C GDAT 10
C ..................................................................GDAT 20
C GDAT 30
C SUBROUTINE GDATA GDAT 40
C GDAT 50
C PURPOSE GDAT 60
C GENERATE INDEPENDENT VARIABLES UP TO THE M-TH POWER (THE GDAT 70
C HIGHEST DEGREE POLYNOMIAL SPECIFIED) AND COMPUTE MEANS, GDAT 80
C STANDARD DEVIATIONS, AND CORRELATION COEFFICIENTS. THIS GDAT 90
C SUBROUTINE IS NORMALLY CALLED BEFORE SUBROUTINES ORDER, GDAT 100
C MINV AND MULTR IN THE PERFORMANCE OF A POLYNOMIAL GDAT 110
C REGRESSION. GDAT 120
C GDAT 130
C USAGE GDAT 140
C CALL GDATA (N,M,X,XBAR,STD,D,SUMSQ) GDAT 150
C GDAT 160
C DESCRIPTION OF PARAMETERS GDAT 170
C N - NUMBER OF OBSERVATIONS. GDAT 180
C M - THE HIGHEST DEGREE POLYNOMIAL TO BE FITTED. GDAT 190
C X - INPUT MATRIX (N BY M+1) . WHEN THE SUBROUTINE IS GDAT 200
C CALLED, DATA FOR THE INDEPENDENT VARIABLE ARE GDAT 210
C STORED IN THE FIRST COLUMN OF MATRIX X, AND DATA FORGDAT 220
C THE DEPENDENT VARIABLE ARE STORED IN THE LAST GDAT 230
C COLUMN OF THE MATRIX. UPON RETURNING TO THE GDAT 240
C CALLING ROUTINE, GENERATED POWERS OF THE INDEPENDENTGDAT 250
C VARIABLE ARE STORED IN COLUMNS 2 THROUGH M. GDAT 260
C XBAR - OUTPUT VECTOR OF LENGTH M+1 CONTAINING MEANS OF GDAT 270
C INDEPENDENT AND DEPENDENT VARIABLES. GDAT 280
C STD - OUTPUT VECTOR OF LENGTH M+1 CONTAINING STANDARD GDAT 290
C DEVIATIONS OF INDEPENDENT AND DEPENDENT VARIABLES. GDAT 300
C D - OUTPUT MATRIX (ONLY UPPER TRIANGULAR PORTION OF THE GDAT 310
C SYMMETRIC MATRIX OF M+1 BY M+1) CONTAINING CORRELA- GDAT 320
C TION COEFFICIENTS. (STORAGE MODE OF 1) GDAT 330
C SUMSQ - OUTPUT VECTOR OF LENGTH M+1 CONTAINING SUMS OF GDAT 340
C PRODUCTS OF DEVIATIONS FROM MEANS OF INDEPENDENT GDAT 350
C AND DEPENDENT VARIABLES. GDAT 360
C GDAT 370
C REMARKS GDAT 380
C N MUST BE GREATER THAN M+1. GDAT 390
C IF M IS EQUAL TO 5 OR GREATER, SINGLE PRECISION MAY NOT BE GDAT 400
C SUFFICIENT TO GIVE SATISFACTORY COMPUTATIONAL RESULTS. GDAT 410
C GDAT 420
C SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED GDAT 430
C NONE GDAT 440
C GDAT 450
C METHOD GDAT 460
C REFER TO B. OSTLE, 'STATISTICS IN RESEARCH', THE IOWA STATE GDAT 470
C COLLEGE PRESS, 1954, CHAPTER 6. GDAT 480
C GDAT 490
C ..................................................................GDAT 500
C GDAT 510
SUBROUTINE GDATA (N,M,X,XBAR,STD,D,SUMSQ) GDAT 520
DIMENSION X(1),XBAR(1),STD(1),D(1),SUMSQ(1) GDAT 530
C GDAT 540
C ...............................................................GDAT 550
C GDAT 560
C IF A DOUBLE PRECISION VERSION OF THIS ROUTINE IS DESIRED, THE GDAT 570
C C IN COLUMN 1 SHOULD BE REMOVED FROM THE DOUBLE PRECISION GDAT 580
C STATEMENT WHICH FOLLOWS. GDAT 590
C GDAT 600
C DOUBLE PRECISION X,XBAR,STD,D,SUMSQ,T1,T2 GDAT 610
C GDAT 620
C THE C MUST ALSO BE REMOVED FROM DOUBLE PRECISION STATEMENTS GDAT 630
C APPEARING IN OTHER ROUTINES USED IN CONJUNCTION WITH THIS GDAT 640
C ROUTINE. GDAT 650
C GDAT 660
C THE DOUBLE PRECISION VERSION OF THIS SUBROUTINE MUST ALSO GDAT 670
C CONTAIN DOUBLE PRECISION FORTRAN FUNCTIONS. SQRT AND ABS IN GDAT 680
C STATEMENT 180 MUST BE CHANGED TO DSQRT AND DABS. GDAT 690
C GDAT 700
C ...............................................................GDAT 710
C GDAT 720
C GENERATE INDEPENDENT VARIABLES GDAT 730
C GDAT 740
IF(M-1) 105, 105, 90 GDAT 750
90 L1=0 GDAT 760
DO 100 I=2,M GDAT 770
L1=L1+N GDAT 780
DO 100 J=1,N GDAT 790
L=L1+J GDAT 800
K=L-N GDAT 810
100 X(L)=X(K)*X(J) GDAT 820
C GDAT 830
C CALCULATE MEANS GDAT 840
C GDAT 850
105 MM=M+1 GDAT 860
DF=N GDAT 870
L=0 GDAT 880
DO 115 I=1,MM GDAT 890
XBAR(I)=0.0 GDAT 900
DO 110 J=1,N GDAT 910
L=L+1 GDAT 920
110 XBAR(I)=XBAR(I)+X(L) GDAT 930
115 XBAR(I)=XBAR(I)/DF GDAT 940
C GDAT 950
DO 130 I=1,MM GDAT 960
130 STD(I)=0.0 GDAT 970
C GDAT 980
C CALCULATE SUMS OF CROSS-PRODUCTS OF DEVIATIONS GDAT 990
C GDAT1000
L=((MM+1)*MM)/2 GDAT1010
DO 150 I=1,L GDAT1020
150 D(I)=0.0 GDAT1030
DO 170 K=1,N GDAT1040
L=0 GDAT1050
DO 170 J=1,MM GDAT1060
L2=N*(J-1)+K GDAT1070
T2=X(L2)-XBAR(J) GDAT1080
STD(J)=STD(J)+T2 GDAT1090
DO 170 I=1,J GDAT1100
L1=N*(I-1)+K GDAT1110
T1=X(L1)-XBAR(I) GDAT1120
L=L+1 GDAT1130
170 D(L)=D(L)+T1*T2 GDAT1140
L=0 GDAT1150
DO 175 J=1,MM GDAT1160
DO 175 I=1,J GDAT1170
L=L+1 GDAT1180
175 D(L)=D(L)-STD(I)*STD(J)/DF GDAT1190
L=0 GDAT1200
DO 180 I=1,MM GDAT1210
L=L+I GDAT1220
SUMSQ(I)=D(L) GDAT1230
180 STD(I)= SQRT( ABS(D(L))) GDAT1240
C GDAT1250
C CALCULATE CORRELATION COEFFICIENTS GDAT1260
C GDAT1270
L=0 GDAT1280
DO 190 J=1,MM GDAT1290
DO 190 I=1,J GDAT1300
L=L+1 GDAT1310
190 D(L)=D(L)/(STD(I)*STD(J)) GDAT1320
C GDAT1330
C CALCULATE STANDARD DEVIATIONS GDAT1340
C GDAT1350
DF=SQRT(DF-1.0) GDAT1360
DO 200 I=1,MM GDAT1370
200 STD(I)=STD(I)/DF GDAT1380
RETURN GDAT1390
END GDAT1400