def main():
""" Runs independently from command line to test the GALE algorithm. """
graphPerformance = False # Built in graphing ability, currently not functional, but mechanism is in place.
trainData = "2_1000_0_1600_0_0_CV_0_Train.txt"
testData = "2_1000_0_1600_0_0_CV_0_Test.txt"
outProg = "GH_GALE_ProgressTrack"
outPop = "GH_GALE_PopulationOut"
bitLength = 1 # This implementation is not yet set up to handle other rule representations, or bit encoding lengths.
CVpartitions = 10
trackCycles = 1
iterInput = '5.10.20'
xdim = 10
ydim = 10
dist = 2
wild = 0.75
prune = 1
#Figure out the iteration stops for evaluation, and the max iterations.
iterList = iterInput.split('.')
for i in range(len(iterList)):
iterList[i] = int(iterList[i])
lastIter = iterList[len(iterList)-1]
#Sets up up algorithm to be run.
GALEConstants.setConstants(prune, wild)
e = GALE_Environment(trainData,testData,bitLength)
sampleSize = e.getNrSamples()
gale = GALE(e, outProg, outPop, bitLength, CVpartitions, graphPerformance, xdim, ydim, dist)
#Set some GALE parameters.
if trackCycles == 'Default':
gale.setTrackingIterations(sampleSize)
else:
gale.setTrackingIterations(trackCycles)
gale.setNumberOfTrials(lastIter, iterList)
#Run the GALE Algorithm
gale.runGALE()
12: Pruning - '0' or '1'
13: Wild frequency - 0.5, 0.75
"""
# *************************************************************************************************
graphPerformance = False # Built in graphing ability, currently not functional, but mechanism is in place. NOT MEANT TO BE USED ON CLUSTER.
numArgs = len(argv)
print "Arguments: " + str(numArgs)
if numArgs == 14:
if argv[1] == 'gh': #Different rule representations could be programmed but have not been in this implementation.
print ("Format Training data: "+argv[2]+" using a "+argv[6]+" bit coding scheme.")
#Sets up up algorithm to be run.
GALEConstants.setConstants(int(argv[12]), float(argv[13]))
e = GALE_Environment(str(argv[2]), str(argv[3]), int(argv[6]))
sampleSize = e.getNrSamples()
gale = GALE(e, argv[4], argv[5], int(argv[6]), int(argv[7]), graphPerformance, int(argv[10]), int(argv[10]), int(argv[11]))
#Figure out the iteration stops for evaluation, and the max iterations.
iterList = argv[9].split('.')
for i in range(len(iterList)):
iterList[i] = int(iterList[i])
lastIter = iterList[len(iterList)-1]
#Set some GALE parameters.
if argv[9] == 'Default':
gale.setTrackingIterations(sampleSize)
else:
gale.setTrackingIterations(int(argv[8]))
gale.setNumberOfTrials(lastIter, iterList)