def __init__(self, kind_string):
"""テーブルのコンストラクタ。"""
print "[table]テーブルのコンストラクタの起動"
self._kind = kind_string
self._attributes = attributes.Attributes(self._kind)
self._images = []
self._smallImages = []
self._tuples = []
return
def __init__(self, kind_string):
"""テーブルのコンストラクタ。"""
self._kind_string = kind_string
self._attributes = attributes.Attributes(kind_string)
self._images = []
self._thumbnails = []
self._tuples = []
return
def __init__(self, classifier, training_data, attribute_set):
self.classifier = classifier
self.training_data = training_data
self.attribute_set = attributes.Attributes(
False, sorted(attribute_set, key=lambda attribute: attribute.name))
self.root = None
self.build_decision_tree()
return
dest='attributes_file',
required=True)
parser.add_argument('--train',
type=argparse.FileType('r'),
help='Name of the file to use for training',
dest='training_file',
required=True)
parser.add_argument('--test',
type=argparse.FileType('r'),
dest='testing_file',
help='Name of the file to use for testing')
args = parser.parse_args()
# Read in a complete list of attributes.
# global all_attributes
all_attributes = attributes.Attributes(args.attributes_file)
if args.classifier not in all_attributes.all_names():
sys.stderr.write("Classifier '%s' not a recognized attribute name\n" %
args.classifier)
sys.exit(1)
classifier = all_attributes[args.classifier]
# Import the d-tree module, removing the .py extension if found
if args.dtree_module.endswith('.py') and len(args.dtree_module) > 3:
dtree_pkg = __import__(args.dtree_module[:-3])
else:
dtree_pkg = __import__(args.dtree_module)
# Train
training_data = dataset.DataSet(args.training_file, all_attributes)
starting_attrs = copy.copy(all_attributes)
def mainscript():
# parse argument
ap = argparse.ArgumentParser(
description='Handle Missing Values in Dataset')
ap.add_argument(
'--datafile',
'-d',
# type = argparse.FileType('r'),
help='Name of the data file',
dest='datafile',
required=True)
ap.add_argument(
'--testfile',
'-t',
# type=argparse.FileType('r'),
help='Name of the test file',
dest='testfile')
ap.add_argument(
'--attributes',
'-a',
# type=argparse.FileType('r'),
help='Name of the attribute specification file',
dest='attributes_file',
required=True)
# ap.add_argument('--intermediate', '-inter',
# help='Name of attribute intermeiate output file',
# dest='inter'
# )
ap.add_argument('--output',
'-oa',
type=argparse.FileType('w'),
help='Name of attribute output file',
dest='att_outfile',
default=sys.stdout)
ap.add_argument('--output2',
'-od',
type=argparse.FileType('w'),
help='Name of training set output file',
dest='train_outfile',
default=sys.stdout)
ap.add_argument('--output3',
'-ot',
type=argparse.FileType('w'),
help='Name of test set output file',
dest='test_outfile',
default=sys.stdout)
args = ap.parse_args([
'--datafile', './dataset/src/adult.dat', '--attributes',
'./dataset/src/adult_attrib.txt', '--testfile',
'./dataset/src/adult_test.dat', '-inter',
'./dataset/preprocessed/adult_attrib.txt', '-oa',
'./dataset/processed/adult_attrib.txt', '-od',
'./dataset/processed/adult.csv', '-ot',
'./dataset/processed/adult_test.csv'
])
# create training datatable
preprocess_file(args.datafile, args.datafile + '_preprocessed')
datatable = create_data_table(args.datafile + '_preprocessed')
# create testing datatable
if args.testfile:
preprocess_file(args.testfile, args.testfile + '_preprocessed')
testtable = create_data_table(args.datafile + '_preprocessed')
# create attribute file
preprocess_file(args.attributes_file,
args.attributes_file + '_preprocessed')
file = open(args.attributes_file + '_preprocessed', 'r')
all_attributes = attributes.Attributes(file)
file.close()
mod_attributes = modify_attributes(datatable, all_attributes)
if testtable:
mod_attributes = modify_attributes(testtable, mod_attributes)
# output attibutes file
output_attribute(mod_attributes, args.att_outfile)
# output datatable
output_datatable(datatable, args.train_outfile)
output_datatable(datatable, args.test_outfile)
def parse_attributes(character_info):
attribute_info = dict()
for key, value in character_info["Attributes"].items():
attribute_info[key.lower().replace(' ', '_')] = int(value)
return attributes.Attributes(**attribute_info)
