Orange Forum • View topic - Controlling nomogram appearance

Controlling nomogram appearance

A place to ask questions about methods in Orange and how they are used and other general support.

Controlling nomogram appearance

Postby bgbg » Tue Jun 23, 2009 12:04

Hello,
I have produced a nomogram of bayes classifier. However, I need to bring it to the so-called "publication quality". Mainly, I need to enlarge the font size, and make much thicker lines. The following code attempts to do exactly this task, but it has no effect on the graph I see on the screen:

Code: Select all
import sys
import orange, orngSVM, orngBayes, orngTest, orngStat
import os
sys.path.append(r"c:\Python25\lib\site-packages\orange\OrangeWidgets\Classify")
import OWROC
import OWNomogram
import pickle


strDir = r'Q:\some\dir'
strDataFile = r'data.dump'

data = pickle.load(open(os.path.join(strDir, strDataFile)))
b = orngBayes.BayesLearner()
b.name = 'bayes'
learners = [b]
results = orngTest.learnAndTestOnLearnData(learners, data)
a = OWNomogram.QApplication(sys.argv)
ow = OWNomogram.OWNomogram()
ow.lineWidth = 2
ow.alignType = 1 # by zero influence
ow.contType = 2 # 0 - 1D, 1 - 2D
ow.yAxis = 0
ow.fontSize = 12
ow.sort_type = 2 # positive influence
ow.TargetClassIndex = 1
ow.classifier(b(data))
   
ow.show()
a.exec_loop()


Any ideas?

Return to Questions & Support