def main():
# define the question # and question column set
problem = 2
csv = 'project510Data.csv'
dataTotal = dTools.dataTotal(csv)
q = qTools.Question(problem)
q.getInputDataAndLabels(dataTotal._dataList)
dataQ = dTools.dataQuestion(q, 0.90)
nRuns = 10
allRuns = []
print "----"
for ratio in range(950, 1000, 50):
dataQ = dTools.dataQuestion(q, 1.0 * ratio / 1000)
for maxIter in range(100, 300, 25):
print "RATIO:", 1.0 * ratio / 1000
print "MAX ITER: %d" % (maxIter)
classifier = aiTools.classifier(q, dataQ, aiTools.featureExtractor,
maxIter)
allRuns.append((avgNruns(dataQ, classifier,
nRuns), 1.0 * ratio / 1000, maxIter))
print "--"
print "----"
allRuns.sort()
print "TOP 15 RUNS:"
for x in allRuns[-14:][::-1]:
print x
def main(problem, dataTotal):
import questionTools as qTools
import dataTools as dTools
import aiTools
q = qTools.Question(problem)
q.getInputDataAndLabels(dataTotal._dataList)
dataQ = dTools.dataQuestion(q, 0.90)
classifier = aiTools.classifier(q, dataQ, aiTools.featureExtractor, 200)
manager = multiprocessing.Manager()
return_dict = manager.dict()
tasks = manager.Queue()
num_processes = 10
pool = multiprocessing.Pool(num_processes)
proc = []
for i in range(num_processes):
process_name = 'Proc%i' % i
p = multiprocessing.Process(target=avgNruns,
args=(dataQ, classifier, problem,
return_dict, i))
proc.append(p)
p.start()
for p in proc:
p.join()
runs = return_dict.values()
avg = 1.0 * sum(runs) / len(runs)
print "\nPROBLEM %d\nOver %d runs, accuracy averaged: %.2f%s\n" % (
problem, len(runs), avg * 100, "%")
def main(problem, dataTotal):
import questionTools as qTools
import dataTools as dTools
import aiTools
q = qTools.Question(problem)
q.getInputDataAndLabels(dataTotal._dataList)
num_processes = 5
allRuns = []
for RATIO in range(35, 60, 5):
dataQ = dTools.dataQuestion(q, RATIO / 100.0)
for MAX_ITER in range(20, 80, 5):
classifier = aiTools.classifier(q, dataQ, aiTools.featureExtractor,
MAX_ITER)
manager = multiprocessing.Manager()
return_dict = manager.dict()
tasks = manager.Queue()
pool = multiprocessing.Pool(num_processes)
proc = []
for i in range(num_processes):
process_name = 'Proc%i' % i
p = multiprocessing.Process(target=avgNruns,
args=(dataQ, classifier, problem,
return_dict, i))
proc.append(p)
p.start()
for p in proc:
p.join()
runs = return_dict.values()
avg = 1.0 * sum(runs) / len(runs)
print "\nPROBLEM %d\t%d\t%.2f\nOver %d runs, accuracy averaged: %.2f%s\n" % (
problem, MAX_ITER, RATIO / 100.0, len(runs), avg * 100, "%")
allRuns.append((avg, MAX_ITER, RATIO / 100.0))
allRuns.sort()
allRuns = allRuns[-14:]
del classifier
del manager, return_dict, p, proc, runs
del dataQ
print "TOP 15 RUNS:"
for x in allRuns[::-1]:
print x
def main():
# define the question # and question column set
problem = 2
csv = 'project510Data.csv'
dataTotal = dTools.dataTotal(csv)
q = qTools.Question(problem)
q.getInputDataAndLabels(dataTotal._dataList)
trainToTestRatio = 0.9
dataQ = dTools.dataQuestion(q, trainToTestRatio)
classifier = aiTools.classifier(q, dataQ, aiTools.featureExtractor,
maxIter)
classifier.prepDataTrain()
classifier.train()
tests.append(classifier.test())
def main():
# define the question # and question column set
problem = 2
csv = 'project510Data.csv'
dataTotal = dTools.dataTotal(csv)
q = qTools.Question(problem)
q.getInputDataAndLabels(dataTotal._dataList)
trainToTestRatio = 0.9
dataQ = dTools.dataQuestion(q, trainToTestRatio)
nRuns = 10
classifier = aiTools.classifier(q, dataQ, aiTools.featureExtractor,
maxIter)
avgNruns(dataQ, classifier, nRuns)