Google
 

Trailing-Edge - PDP-10 Archives - decuslib20-01 - decus/20-0025/stades.txt
There is 1 other file named stades.txt in the archive. Click here to see a list.
00010	THESE ARE INSTRUCTIONS FOR THE FOLLOWING STATISTICAL PROGRAMS:
00013	   PROGRAM                PURPOSE
00015	   UNISTA*    DESCRIPTION OF UNIVARIATE DATA
00016	   CURFIT*    FITS SIX DIFFERENT CURVES (LEAST SQUARES)
00017	   MULFIT*    MULTIPLE LINEAR FIT - WITH TRANSFORMATIONS
00018	   COLINR*    CONFIDENCE LIMITS ON SIMPLE LINEAR REGRESSION
00020	THE INPUT FORMAT IS IDENTICAL FOR THESE FOUR PROGRAMS.  THE 
00021	LINE NUMBERS 1-699 ARE FREE FOR USE AS DATA STATEMENTS, AS
00022	REQUIRED.  ALL VALUES FOR ONE VARIABLE ARE GIVEN IN DATA
00023	STATEMENTS, FOLLOWED BY ALL DATA FOR A SECOND VARIABLE, AND SO
00024	ON. ONE CONVENIENT WAY TO ARRANGE INPUT IS TO PUT ALL VALUES OF
00025	VARIABLE 1 AT LINES 100-199, ALL FOR VARIABLE 2 AT LINES 200-299,
00026	ETC.
00029	UNISTA ACCEPTS UP TO 300 OBSERVATIONS ON ONE VARIABLE.
00031	CURFIT ACCEPTS UP TO 200 OBSERVATIONS ON TWO VARIABLES.  THE
00032	  DATA FOR THE DEPENDENT VARIABLE (Y) IS SUPPLIED FIRST, FOLLOW-
00033	 ED BY  THE DATA FOR THE INDEPENDENT VARIABLE (X).
00038	MULFIT ACCEPTS UP TO 61 OBSERVATIONS ON UP TO 6 VARIABLES.  THE
00039	  ORDER OF THE VARIABLES IN THE INPUT IS IMPORTANT ONLY IF
00040	  TRANSFORMATIONS ARE NOT BEING USED.  IF NO TRANSFORMED VAR-
00041	  IABLES ARE GIVEN, THE PROGRAM WILL FIT A LINEAR FUNCTION
00042	  FOR THE FIRST DATA VARIABLE IN TERMS OF THE FOLLOWING VARI-
00043	  ABLES.  BY USING TRANSFORMS, ANY FUNCTION OF ANY OR ALL OF
00044	  THE INPUT VARIABLES MAY BE USED AS ANY VARIABLE IN THE ANALYSIS.
00048	      10XX  LET X(J)= ANY FUNCTION OF V(1), V(2),...., V(R)
00050	  WHERE X(J) IS THE J-TH VARIABLE IN THE ANALYSIS (J=1 IS THE
00051	  DEPENDENT VARIABLE), AND THE V(I) ARE THE INPUT DATA VARIABLES.
00052	  THUS, ONE MIGHT ENTER DATA FOR THREE VARIABLES V(1), V(2),
00053	  AND V(3), BUT RUN THE ANALYSIS WITH X(1)=V(2) AS THE DEPENDENT
00054	  VARIABLE, AND THREE 'INDEPENDENT' VARIABLES, BY ENTERING 
00055	  TRANSFORMATIONS LIKE THE FOLLOWING AT LINES 1000-1003:
00057	      1000 LET X(1)=V(2)
00058	      1001 LET X(2)=LOG(V(1)+V(3))
00059	      1002 LET X(3)=V(1)+X(2)*EXP(-1-V(3))
00060	      1003 LET X(4)=(V(3)-1)^2
00062	  NOTE: IF TRANSFORMS ARE USED, THE PROGRAM EXPECTS AT LEAST
00063	  THREE SETS OF INPUT VARIABLES. IF THERE ARE ONLY TWO SETS OF
00064	  VALID DATA, A SET OF ZEROS MAY BE INPUT FOR THE THIRD VARIABLE.
00065	  HOWEVER, THIS DUMMY SET SHOULD NOT BE USED IN DEFINING THE
00066	  TRANSFORMATIONS.
00067	  
00068	COLINR ACCEPTS UP TO 250 OBSERVATIONS ON TWO VARIABLES.
00069	  IT ASSUMES THAT THE FIRST DATA VARIABLE IS THE DEPENDENT ONE,
00070	  AND ALSO ALLOWS FOR THE ADDITION OF ADDITIONAL VALUES OF THE
00071	  INDEPENDENT VARIABLE IN THE DATA LIST, AS IN CURFIT.
00073	END OF STADES*** EXPLANATION