source: orange/docs/tutorial/rst/code/domain7.py @ 9746:f2ff17dfd86e

Revision 9746:f2ff17dfd86e, 1005 bytes checked in by Miha Stajdohar <miha.stajdohar@…>, 2 years ago (diff)

Changed data set path.

Line 
1# Description: Shows how to add class noise and missing attributes to data sets. Also shows how to test a single learner on a range of data sets.
2# Category:    preprocessing
3# Uses:        imports-85
4# Referenced:  domain.htm
5
6import orange
7
8def report_prob(header, data):
9  print 'Size of %s: %i instances; ' % (header, len(data)),
10  n = 0
11  for i in data:
12    if int(i.getclass()) == 0:
13      n = n + 1
14  if len(data):
15    print "p(%s)=%5.3f" % (data.domain.classVar.values[0], float(n) / len(data))
16  else:
17    print
18
19filename = "adult_sample.tab"
20data = orange.ExampleTable(filename)
21report_prob('data', data)
22
23selection = [1] * 10 + [0] * (len(data) - 10)
24data1 = data.select(selection)
25report_prob('data1, first ten instances', data1)
26
27data2 = data.select(selection, negate=1)
28report_prob('data2, other than first ten instances', data2)
29
30selection = [1] * 12 + [2] * 12 + [3] * 12 + [0] * (len(data) - 12 * 3)
31data3 = data.select(selection, 3)
32report_prob('data3, third dozen of instances', data3)
Note: See TracBrowser for help on using the repository browser.