| title 'ppseps.sas: PPS and EPS, 10 firms, p 335'; options linesize=90; |
Explanation |
| ods html4 path="~/public_html" file="ppseps.html"; | Explanation |
|
Data Preparation |
| data ppseps; infile "ppseps.csv" dsd dlm=","; length firm$ 23; input firm$ eps pps; |
Explanation Note: To calculate a forecast for a value of X not in the data set, add a new data record to the data file with the desired value of X. Set the corresponding value of Y as missing, with a decimal point with no numbers on either side of it. |
|
Core Analyses |
| proc reg data=ppseps; model pps = eps / clb clm cli; |
Regress pps onto eps with the following options: calculate the corresponding confidence intervals of the b's as well as the confidence intervals about the mean predictions and the prediction intervals for individual forecasts. |
|
Hi-Resolution Graphics |
| filename gfile 'ppseps.pdf'; goptions device=pdfc gsfmode=replace gsfname=gfile rotate=landscape; |
Explanation with the addition of a landscape display instead of the default portrait display. |
| symbol1 value=dot cv=blue width=5; symbol2 interpol=rlclm95 value=none c=green width=4; symbol3 interpol=rlcli95 value=none c=red width=4; |
Define three separate sets of symbols for three separate graphic analysis, such as for the subsequent Proc Gplot. The first set of symbols defines the points in the scatterplot. The second set specifies the regression line and defines the color and width of the 95% confidence intervals for the mean of the predicted values (rlclm95). The third set specifies the regression line and defines the color and width of the 95% prediction intervals for individual predictons (rlcli95). |
| proc gplot data=ppseps; plot pps * (eps eps eps) / overlay; title 'Regression for PPS and EPS'; |
Run three separate plots of pps versus eps. Overlay all three plots on the same graph. The respective symbol statements from above specify the characteristics of each plot. The result is a single graph that displays the scatterplot, the regression line, and the associated confidence and prediction intervals. |