Regression problems in a Lin/Log graph

From: Bill ?sterlund (billiboj_at_yahoo.se)
Date: 10/08/04


Date: 7 Oct 2004 22:58:32 -0700

Hi All!

I have created a Lin/Log graph where the Y-axis is logarithmic from
approx 0.1 to 10, and the X-axis ranges from approx 0% to 100% and is
logarithmic up to 50% and then reversed, so it is symmetric with 50%
in the middle. In that graph is a regression line, it should be
straight and heading for the coordinates 0,0 (or approx 0,0). An
ordinary regression line will not be straight but get flat at high
percents. From that line I need to caclulate a couple of figures on
the Y axis. This is what my customers are asking for.

What I need is to know to accomplish that, is how do I calculate an...
 ...X value when I know the Y value?
 ...Y value when I know the X value?

If you got the time, since I am no statistician but a developer,
please be as detailed as you possible. But just a hint how to move on
is appreciciated as well.
 
For thosee of you who knows SAS software: I have produced the graph
using the code below (note the values on the axes).

data temp;
 input YVAR XVAR;
 cards;
 0.3 0.045
 0.4 0.07
 0.75 0.24
 1.1 0.14
 2.0 0.39
 3.0 0.66
 5 0.875
 ;
run;

data annoplot;
 xsys = '4';
 ysys = '2';
 function = 'label';
  x=3; y=0.32; text = '0.3'; output;
 function = 'label';
  x=3; y=0.45; text = '0.4'; output;
 function = 'label';
  x=3; y=0.81; text = '0.75'; output;
 function = 'label';
  x=3; y=1.2; text = '1.10'; output;
 function = 'label';
  x=3; y=2.2; text = '2.00'; output;
 function = 'label';
  x=3; y=3.3; text = '3.00'; output;
 function = 'label';
  x=3; y=5.5; text = '5.00'; output;
 xsys = '2';
 ysys = '2';
 function = 'move';
  x=0.0025; y=0.15; output;
 function = 'draw';
  x=0.95; y=8; output;
run;

symbol1 value=dot height=3 h=1.5;

 axis1 logstyle=expand logbase=10 label=(' ')
        major=(height=1.5) minor=( height=1) ;

 axis2 logstyle=expand logbase=10 label=(font = 'courier new' h=1.3
'Blah blah blah...')
  order=(0.0001,0.001,0.01,0.05,0.10,0.25,0.50,0.75,0.90,0.95,0.99,0.9998,0.9999)
     value=("0.01" "0.1" "1" "5" "10" "25" "50" "75" "90" "95" "99"
"99.98" "99.99");

proc gplot data=temp;
    plot YVAR * XVAR /
      LHREF=33
      vref = 0.3, 0.4, 0.75, 1.10, 2.00, 3.00, 5.00
      vaxis=axis1
      haxis=axis2
      anno=annoplot
 ;
 title1 font = 'courier new' h=1.3 JUSTIFY=LEFT "Report No: 1 ";
 title2 ' ';
 title3 ' ';
 title4 font = 'courier new' h=1.3 JUSTIFY=LEFT "Blah blah blah...";
run;
quit;

TIA
/Bill



Relevant Pages

  • Regression problems in a Lin/Log graph
    ... I have created a Lin/Log graph where the Y-axis is logarithmic from ... straight and heading for the coordinates 0,0. ... 'Blah blah blah...') ...
    (sci.stat.edu)
  • Re: Using Graph properties in VBA
    ... Dim iPts as long ... the graph to the appropriate colours, but the issue now is that there were ... ins and outs when I wrote the macro, ...
    (microsoft.public.excel.programming)
  • Re: how to reduce this signal flow graph?
    ... so your graph looks like depicted below, with summing nodes ... and each edge labeled d = 1/, have I got that straight? ... There are two parallel paths from node 1 to node 0. ...
    (sci.math)
  • Lines of a particular slope that can be moved around
    ... I am plotting a graph that can be approximated using 2 straight lines ... The slope of the 2 lines (approximating the ... user has to change the slope of the line each time parameters change. ...
    (microsoft.public.excel.charting)
  • Re: How do I create a curved graph in excel?
    ... David Biddulph ... graph line connecting the data points is a straight line. ...
    (microsoft.public.excel.charting)

Quantcast