def main():
# get training file, test file and output file's name from command line argument
train_file = sys.argv[1]
test_file = sys.argv[2]
output_file = sys.argv[3]
# get training file and make list of columns
with open(train_file) as f:
train_data = f.readlines()
train_data = [d.strip() for d in train_data]
# make train_data string to attribute list and get each attribute name and total number of columns
attribute_list = []
total_attribute = 0
for line in train_data:
each_line = line.split("\t")
attribute_list.append(each_line)
total_attribute += 1
attribute_name = attribute_list[0]
attribute_list.pop(0)
total_attribute -= 1
possible_name = {}
# get possible classes
for i in range(len(attribute_name)):
ith_class = []
for attr in attribute_list:
if attr[i] not in ith_class:
ith_class.append(attr[i])
possible_name[attribute_name[i]] = ith_class
# get decision class's label and generate tree
decision_label = attribute_name[len(attribute_name) - 1]
tree = DecisionTree.GenerateTree(attribute_list, attribute_name)
# get test input from test file
with open(test_file) as f:
test_input = f.readlines()
test_input = [d.strip() for d in test_input]
test_input.pop(0)
test_data = [[]]
for line in test_input:
test_data.append(line.split("\t"))
test_data.remove([])
# predict decision class and write in output file
f = open(output_file, "w")
del attribute_name[attribute_name.index(decision_label)]
for name in attribute_name:
f.write(name + "\t")
f.write(decision_label + "\n")
for entry in test_data:
# call decision tree
tempDict = tree.copy()
parentDict = tempDict
rootDict = tree.copy()
rootDict = rootDict[list(rootDict.keys())[0]]
result = ""
# trace tree while the answer is found or tree is end
while isinstance(tempDict, dict):
root = TreeNode.TreeNode(
list(tempDict.keys())[0], tempDict[list(tempDict.keys())[0]])
tempDict = tempDict[list(tempDict.keys())[0]]
index = attribute_name.index(root.value)
value = entry[index]
if (value in list(tempDict.keys())):
child = TreeNode.TreeNode(value, tempDict[value])
result = tempDict[value]
parentDict = tempDict
tempDict = tempDict[value]
# can't find the entry in tree, follow majority vote
else:
result = getMaxLabel(parentDict, tempDict, rootDict)
break
for i in entry:
f.write(i + "\t")
f.write(result + "\n")
f.close()
