def crossall():
outfile = open('results', 'w', 1)
cmd = "./text-train.py train -f -P "
for i in range(8):
cmd2 = cmd + str(i) + " -F "
for j in range(4):
cmd3 = cmd2 + str(j) + " -N "
for k in range(2):
cmd4 = cmd3 + str(k) + " -L "
for l in range(4):
name = "train"
outfile.write(str(i) + str(j) + str(k) + str(l) + '\n')
cmd5 = cmd4 + str(l) + " " + name + ".model"
outfile.write(cmd5 + '\n')
acc = []
for m in range(3):
fold(m)
call(cmd5)
print cmd5
acc.append(
text_predict.main([
0, "testfile", name + ".model", "out", "-f",
"-a", "0"
]))
outfile.write(str(float(sum(acc) / len(acc))) + '\n')
outfile.flush()
outfile.close()
def main():
outfile = open('final', 'w', 1)
acc = []
cmd = "./libshorttext/text-train.py train -f -P 3 -F 0 -N 1 -L 2"
#cmd = './text-train.py train -f'
#cmd = "./text-train.py train -f -P 3 -F 0 -N 1 -L 2 -A train_feats"
#cmd = "./text-train.py train -f -A train_feats"
outfile.write(cmd+'\n')
confusion_table = None
for m in range(10):
fold(m)
call(cmd)
#acc.append(text_predict.main([0, "testfile", "train.model", "out1", "-f", "-A", "test_feats"]))
acc.append(text_predict.main([0, "testfile", "train.model", "out1", "-f"]))
outfile.write('fold ' + str(m) + ' acc ' + str(acc[m])+'\n')
analyzer = Analyzer('train.model')
insts = InstanceSet('out1')
confusion_table = analyzer.get_confusion_table(insts, confusion_table)
outfile.write('average: ' + str(float(sum(acc) / len(acc))) + '\n')
analyzer.draw_confusion_table(insts, confusion_table, outfile)
def main():
outfile = open('final', 'w', 1)
acc = []
cmd = "./libshorttext/text-train.py train -f -P 3 -F 0 -N 1 -L 2"
#cmd = './text-train.py train -f'
#cmd = "./text-train.py train -f -P 3 -F 0 -N 1 -L 2 -A train_feats"
#cmd = "./text-train.py train -f -A train_feats"
outfile.write(cmd + '\n')
confusion_table = None
for m in range(10):
fold(m)
call(cmd)
#acc.append(text_predict.main([0, "testfile", "train.model", "out1", "-f", "-A", "test_feats"]))
acc.append(
text_predict.main([0, "testfile", "train.model", "out1", "-f"]))
outfile.write('fold ' + str(m) + ' acc ' + str(acc[m]) + '\n')
analyzer = Analyzer('train.model')
insts = InstanceSet('out1')
confusion_table = analyzer.get_confusion_table(insts, confusion_table)
outfile.write('average: ' + str(float(sum(acc) / len(acc))) + '\n')
analyzer.draw_confusion_table(insts, confusion_table, outfile)
def crossall():
outfile = open('results', 'w', 1)
cmd = "./text-train.py train -f -P "
for i in range(8):
cmd2 = cmd + str(i) + " -F "
for j in range(4):
cmd3 = cmd2 + str(j) + " -N "
for k in range(2):
cmd4 = cmd3 + str(k) + " -L "
for l in range(4):
name = "train"
outfile.write(str(i) + str(j) + str(k) + str(l)+'\n')
cmd5 = cmd4 + str(l) + " " + name + ".model"
outfile.write(cmd5+'\n')
acc = []
for m in range(3):
fold(m)
call(cmd5)
print cmd5
acc.append(text_predict.main([0, "testfile", name + ".model", "out", "-f", "-a", "0"]))
outfile.write(str(float(sum(acc)/len(acc)))+'\n')
outfile.flush()
outfile.close()
