文檔0文檔1文檔[[ 2.]]
文檔0文檔2文檔[[ 2.44948974]]
文檔1文檔2文檔[[ 2.44948974]]
## Stop-word filtering 停用詞過濾
CountVectorizer類可以通過設置stop_words參數過濾停用詞,默認是英語常用的停用詞。
from sklearn.feature_extraction.text import CountVectorizer
corpus = [
'UNC played Duke in basketball',
'Duke lost the basketball game',
'I ate a sandwich'
]
vectorizer=CountVectorizer(stop_words='english')
print vectorizer.fit_transform(corpus).todense()
print vectorizer.vocabulary_
輸出結果:
[[0 1 1 0 0 1 0 1]
[0 1 1 1 1 0 0 0]
[1 0 0 0 0 0 1 0]]
{u'duke': 2, u'basketball': 1, u'lost': 4, u'played': 5, u'game': 3, u'sandwich': 6, u'unc': 7, u'ate': 0}
# Stemming and lemmatization 詞根還原和詞形還原
from sklearn.feature_extraction.text import CountVectorizer
corpus = ['He ate the sandwiches',
'Every sandwich was eaten by him']
vectorizer=CountVectorizer(binary=True,stop_words='english')
print vectorizer.fit_transform(corpus).todense()
print vectorizer.vocabulary_
輸出結果:
[[1 0 0 1]
[0 1 1 0]]
{u'sandwich': 2, u'ate': 0, u'sandwiches': 3, u'eaten': 1}
### 讓我們分析一下單詞gathering的詞形還原:
corpus = [
'I am gathering ingredients for the sandwich.',
'There were many wizards at the gathering.'
]
import nltk
nltk.download()
from nltk.stem.wordnet import WordNetLemmatizer
from nltk import word_tokenize
from nltk.stem import PorterStemmer
from nltk.stem.wordnet import WordNetLemmatizer
from nltk import pos_tag
wordnet_tags = ['n', 'v']
corpus = [
'He ate the sandwiches',
'Every sandwich was eaten by him'
]
stemmer = PorterStemmer()
print('Stemmed:', [[stemmer.stem(token) for token in word_tokenize(document)] for document in corpus])
輸出結果:
('Stemmed:', [[u'He', u'ate', u'the', u'sandwich'], [u'Everi', u'sandwich', u'wa', u'eaten', u'by', u'him']])
def lemmatize(token, tag):
if tag[0].lower() in ['n', 'v']:
return lemmatizer.lemmatize(token, tag[0].lower())
return token
lemmatizer = WordNetLemmatizer()
tagged_corpus = [pos_tag(word_tokenize(document)) for document in corpus]
print('Lemmatized:', [[lemmatize(token, tag) for token, tag in document] for document in tagged_corpus])
輸出結果:
('Lemmatized:', [['He', u'eat', 'the', u'sandwich'], ['Every', 'sandwich', u'be', u'eat', 'by', 'him']])
## 帶TF-IDF權重的擴展詞庫
from sklearn.feature_extraction.text import CountVectorizer
corpus=['The dog ate a sandwich, the wizard transfigured a sandwich, and I ate a sandwich']
vectorizer=CountVectorizer(stop_words='english')
print vectorizer.fit_transform(corpus).todense()
print vectorizer.vocabulary_
輸出結果:
[[2 1 3 1 1]]
{u'sandwich': 2, u'wizard': 4, u'dog': 1, u'transfigured': 3, u'ate': 0}
#tf-idf
from sklearn.feature_extraction.text import TfidfVectorizer
corpus = ['The dog ate a sandwich and I ate a sandwich','The wizard transfigured a sandwich']
vectorizer=TfidfVectorizer(stop_words='english')
print vectorizer.fit_transform(corpus).todense()
print vectorizer.vocabulary_
輸出結果:
[[ 0.75458397 0.37729199 0.53689271 0. 0. ]
[ 0. 0. 0.44943642 0.6316672 0.6316672 ]]
{u'sandwich': 2, u'wizard': 4, u'dog': 1, u'transfigured': 3, u'ate': 0}
## 通過哈希技巧實現特征向量
from sklearn.feature_extraction.text import HashingVectorizer
corpus = ['the', 'ate', 'bacon', 'cat']
vectorizer = HashingVectorizer(n_features=6)
print(vectorizer.transform(corpus).todense())
輸出結果:
[[-1. 0. 0. 0. 0. 0.]
[ 0. 0. 0. 1. 0. 0.]
[ 0. 0. 0. 0. -1. 0.]
[ 0. 1. 0. 0. 0. 0.]]
設置成6是為了演示。另外,注意有些單詞頻率是負數。由於Hash碰撞可能發生,所以HashingVectorizer用有符號哈希函數(signed hash function)。特征值和它的詞塊的哈希值帶
同樣符號,如果cats出現過兩次,被哈希成-3,文檔特征向量的第四個元素要減去2。如果dogs出現過兩次,被哈希成3,文檔特征向量的第四個元素要加上2。
## 圖片特征提取
#通過像素值提取特征
scikit-learn的digits數字集包括至少1700種0-9的手寫數字圖像。每個圖像都有8x8像像素構成。每
個像素的值是0-16,白色是0,黑色是16。如下圖所示:
%matplotlib inline
from sklearn import datasets
import matplotlib.pyplot as plt
digits=datasets.load_digits()
print 'Digit:',digits.target[0]
print digits.images[0]
plt.imshow(digits.images[0], cmap=plt.cm.gray_r, interpolation='nearest')
plt.show()
輸出結果:
Digit: 0
[[ 0. 0. 5. 13. 9. 1. 0. 0.]
[ 0. 0. 13. 15. 10. 15. 5. 0.]
[ 0. 3. 15. 2. 0. 11. 8. 0.]
[ 0. 4. 12. 0. 0. 8. 8. 0.]
[ 0. 5. 8. 0. 0. 9. 8. 0.]
[ 0. 4. 11. 0. 1. 12. 7. 0.]
[ 0. 2. 14. 5. 10. 12. 0. 0.]
[ 0. 0. 6. 13. 10. 0. 0. 0.]]

digits=datasets.load_digits()
print('Feature vector:\n',digits.images[0].reshape(-1,64))
輸出結果:
('Feature vector:\n', array([[ 0., 0., 5., 13., 9., 1., 0., 0., 0., 0., 13.,
15., 10., 15., 5., 0., 0., 3., 15., 2., 0., 11.,
8., 0., 0., 4., 12., 0., 0., 8., 8., 0., 0.,
5., 8., 0., 0., 9., 8., 0., 0., 4., 11., 0.,
1., 12., 7., 0., 0., 2., 14., 5., 10., 12., 0.,
0., 0., 0., 6., 13., 10., 0., 0., 0.]]))
%matplotlib inline
import numpy as np
from skimage.feature import corner_harris,corner_peaks
from skimage.color import rgb2gray
import matplotlib.pyplot as plt
import skimage.io as io
from skimage.exposure import equalize_hist
def show_corners(corners,image):
fig=plt.figure()
plt.gray()
plt.imshow(image)
y_corner,x_corner=zip(*corners)
plt.plot(x_corner,y_corner,'or')
plt.xlim(0,image.shape[1])
plt.ylim(image.shape[0],0)
fig.set_size_inches(np.array(fig.get_size_inches())*1.5)
plt.show()