[Jboost-cvs] jboost/scripts AddRandomIndex.py,NONE,1.1

[Yoav F. <yf...@us...>]

Update of /cvsroot/jboost/jboost/scripts
In directory sc8-pr-cvs17.sourceforge.net:/tmp/cvs-serv30612/scripts

Added Files:
	AddRandomIndex.py 
Log Message:
A script for adding a random index to a data file.

--- NEW FILE: AddRandomIndex.py ---
""" 
add an INDEX field to a jboost data file. INDEX is a randomly permuted number ranging
between 1 and the number of examples in the data file. The script also takes care of 
altering the spec file. This pre-processing step makes it possible to track examples through
an n-fold cross validation experiment.
"""

filename = "/Users/yoavfreund/projects/jboost/demo/spambase"

datafile = open(filename+".data",'r')

lines=[]
morelines = datafile.readlines(100000)
while len(morelines)>0:
    lines.extend(morelines)
    morelines = datafile.readlines(100000)
datafile.close()

length = len(lines)

from random import shuffle
shuffle(lines)

newdatafile = open(filename+"I.data",'w')
for i in range(length):
    newdatafile.write(("%d," % i)+lines[i])
    
newdatafile.close()