Username Remember Me?
Password   forgot password?
   
   
Problem with sddecide
Posted: 28 April 2010 10:41 AM   [ Ignore ]  
Member
RankRankRank
Total Posts:  68
Joined  2010-04-27

Hello,

I started to use PRSD and everything goes well excepted the command sddecide which always returns an error. I have tried example 20.3 of the APR course exercises:

====================
a = gendatb(200)
a = sdrelab(a,{1 ‘apple’,2 ‘banana’})
[trtr,trts] = gendat(a,0.5)
w = trtr*nmc
out = trts*w
r = sdroc(out,’confmat’)
fig = sddrawroc(r)
Banana Set, 200 by 2 dataset with 2 classes: [99 101]
1: 1 -> apple
2: 2 -> banana
Banana Set, 200 by 2 dataset with 2 classes: [99 101]
Banana Set, 101 by 2 dataset with 2 classes: [50 51]
Banana Set, 99 by 2 dataset with 2 classes: [49 50]
Nearest Mean, 2 to 2 trained classifier --> affine
Banana Set, 99 by 2 dataset with 2 classes: [49 50]
ROC (96 w-based op.points, 3 measures), curop: 33
est: 1:err(apple)=0.08, 2:err(banana)=0.34, 3:mean-error=0.21

fig =

1

>> sdscatter(ts,sddecide(w,r),’roc’,fig)
??? SWITCH expression must be a scalar or string constant.

Error in ==> /PRSD_Studio_2.1.0_06Apr10_lin32_M75_dev/prsd_toolbox/sddecide.p>sddecide at 56

============================

I have tried on other examples and I still have the same problem.

Any clue of what I’m doing wrong?

thanks in advance

best regards,

Jean-Michel

Profile
 
 
Posted: 28 April 2010 10:47 AM   [ Ignore ]   [ # 1 ]  
Administrator
Avatar
RankRankRankRank
Total Posts:  240
Joined  2008-04-26

Hi Jean-Michel,

that is a bug in sddecide. It was corrected in the latest release 2.1.0 21-apr-2010 (see the release notes: http://perclass.com/doc/guide/intro.html#relnotes )

With Kind Regards,

Pavel

Profile