Code

import os, io
import pandas as pd
import numpy as np
import pickle

from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
from sklearn import metrics
from sklearn.linear_model import LogisticRegressionCV

root = 'C:\Users\xxx\Desktop\SUE\poems'
os.chdir(root)

dirnames = ['\machine','\human']

#contentlist
alldircontent = []
for dirname in dirnames:
    dirpath = root+dirname
    dircontent=[]
    for filename in os.listdir(dirpath):
        fullpath = dirpath + '/' + filename
        with io.open(fullpath, 'r', encoding = 'latin-1') as f:
            text = f.read()
        dircontent.append(text)
    alldircontent.append(dircontent)

#labellist
labels = ['machine','human']
alllabels = []
for i in range(len(alldircontent)):
    lab = labels[i]
    tmp = []
    for elem in alldircontent[i]:
        tmp.append(lab)
    alllabels.append(tmp)

labels_list = [item for sublist in alllabels for item in sublist]
text_list = [item for sublist in alldircontent for item in sublist]

print len(labels_list) == len(text_list)
data = pd.DataFrame()
data['text'] = text_list
data['label'] = labels_list

text_clean=[]
for text in data['text']:
    text = text.rstrip()
    text_clean.append(text)

data['text_clean'] = text_clean
################################################################

ratio = 0.8
mask = np.random.rand(len(data)) < ratio
data_train = data[mask]
data_test = data[~mask]

train_x = data_train['text_clean'].values
train_y = data_train['label'].values

#validation
test_x = data_test['text_clean'].values
test_y = data_test['label'].values

vecSpc = CountVectorizer()
train_feat = vecSpc.fit_transform(train_x)

feat_names = vecSpc.get_feature_names()

clf = LogisticRegressionCV()
clf.fit(train_feat, train_y)

test_feat = vecSpc.transform(test_x)
pred = clf.predict(test_feat)

conf_mat = metrics.confusion_matrix(test_y, pred)
perf_acc = metrics.accuracy_score(test_y, pred)
perf_f1 = metrics.f1_score(test_y, pred, pos_label = 'human')