Thanks a lot for your reply. Actually I am working on a homework for "data mining" class and It looks easy, however I stuck ! I just want to use Data(node)-->Neural network(node)--> sas code(node) I attached what I have. I know it is not practical to ask question this way, but I really appreciate any help you can provide. (How can I send sas data? in attachment?) options nodate nonumber; goptions reset=all gunit=pct; /* Define Graphics Symbols */ symbol1 c=b v=dot h=2 i=none; /* actual model */ symbol2 c=magenta v=circle h=2 i=none; /* observations */ symbol3 c=r v=none i=join; /* predicted model */ AXIS1 OFFSET=(2) length=65 LABEL=(F=Swissb H=3.5 A=90 R=270 "Y") minor=none VALUE=(f=swissb H=3) ORDER = (-6 to 6 by 1) WIDTH=2; AXIS2 OFFSET=(2) LENGTH=85 LABEL=(f=SWISSB H=3.5 "X") minor=none VALUE=(f=swissb H=3) ORDER = (-8 to 8 by 1) WIDTH=2; legend1 across=2 position=(top inside right) value=(f=swissb h=2.5 "#1" "#2") label=none mode=share frame; legend2 across=3 position=(top inside right) value=(f=swissb h=2.5 "True Curve" "Data" "Estimate") label=none mode=share frame; /* True Membership */ title1 h=5 f=swissb 'Figure 1 Two Spirals Problem'; title2 h=4 '(True Membership)'; proc gplot data=&EM_IMPORT_DATA; plot y*x=c / vaxis=axis1 haxis=axis2 legend=legend1 description="plot 1"; run; quit; /* Neural Network fitting with 35 neurons and 50 iterations */ title1 h=5 j=c f=swissb 'Table 1 Two Spirals Problem'; title2 h=4 j=c '(Classification Matrix - Low Iteration)'; proc freq data=&EM_IMPORT_DATA; tables F_c*I_c / norow nocol nopercent missing; run; title1 h=5 f=swissb 'Figure 2 Two Spirals Problem'; title2 h=4 '(Predicted Membership - Low Iteration)'; proc gplot data=&EM_IMPORT_DATA; plot y*x=I_c / vaxis=axis1 haxis=axis2 legend=legend1 description="plot 2"; run; quit;
... View more
