#Success.
import sys, os
from time import time
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn import svm
from sklearn.metrics import classification_report
classes =['pos', 'neg']
train_data =["super and nice", "wonderful","not super", "I don't like", "waste movie", "waste of time","i like it"]
train_labels =['pos','pos','neg','neg','neg','neg','pos']
test_data =["waste, not like it"]
test_labels =["pos"]
print("%d %d" %(len(train_data), len(test_data)))
#min_df(minimum document frequency) & max_df(maximum document frequency)
vectorizer = TfidfVectorizer(min_df=0, max_df =5.0, sublinear_tf=True, use_idf=True)
train_vectors = vectorizer.fit_transform(train_data)
test_vectors = vectorizer.transform(test_data)
#using support vector machine to identify the 
print(len(vectorizer.vocabulary_))
classifier= svm.SVC(kernel='linear')
t0 = time()
classifier.fit(train_vectors, train_labels)
t1 = time()
prediction = classifier.predict(test_vectors)
t2 = time()
time_train = t1-t0
time_predict = t2 - t1
print("Classification Results")
print("%d %d" %(time_train, time_predict))
print(test_labels)
print(prediction)
------
7 1
12
Classification Results
0 0
['pos']
['neg']