NLP教程笔记
目录
TF-IDF算法步骤
第一步，计算词频：
考虑到文章有长短之分，为了便于不同文章的比较，进行"词频"标准化。
第二步，计算逆文档频率：
这时，需要一个语料库（corpus），用来模拟语言的使用环境。
如果一个词越常见，那么分母就越大，逆文档频率就越小越接近0。分母之所以要加1，是为了避免分母为0（即所有文档都不包含该词）。log表示对得到的值取对数。
第三步，计算TF-IDF：
可以看到，TF-IDF与一个词在文档中的出现次数成正比，与该词在整个语言中的出现次数成反比。所以，自动提取关键词的算法就很清楚了，就是计算出文档的每个词的TF-IDF值，然后按降序排列，取排在最前面的几个词。
一、直接实现
import numpy as np
from collections import Counter
import itertools
from visual import show_tfidf # this refers to visual.py in my [repo](https://github.com/MorvanZhou/NLP-Tutorials/)
docs = [
"it is a good day, I like to stay here",
"I am happy to be here",
"I am bob",
"it is sunny today",
"I have a party today",
"it is a dog and that is a cat",
"there are dog and cat on the tree",
"I study hard this morning",
"today is a good day",
"tomorrow will be a good day",
"I like coffee, I like book and I like apple",
"I do not like it",
"I am kitty, I like bob",
"I do not care who like bob, but I like kitty",
"It is coffee time, bring your cup",
]
docs_words = [d.replace(",", "").split(" ") for d in docs]
vocab = set(itertools.chain(*docs_words))
v2i = {v: i for i, v in enumerate(vocab)}
i2v = {i: v for v, i in v2i.items()}
def safe_log(x):
mask = x != 0
x[mask] = np.log(x[mask])
return x
tf_methods = {
"log": lambda x: np.log(1+x),
"augmented": lambda x: 0.5 + 0.5 * x / np.max(x, axis=1, keepdims=True),
"boolean": lambda x: np.minimum(x, 1),
"log_avg": lambda x: (1 + safe_log(x)) / (1 + safe_log(np.mean(x, axis=1, keepdims=True))),
}
idf_methods = {
"log": lambda x: 1 + np.log(len(docs) / (x+1)),
"prob": lambda x: np.maximum(0, np.log((len(docs) - x) / (x+1))),
"len_norm": lambda x: x / (np.sum(np.square(x))+1),
}
def get_tf(method="log"):
# term frequency: how frequent a word appears in a doc
_tf = np.zeros((len(vocab), len(docs)), dtype=np.float64) # [n_vocab, n_doc]
for i, d in enumerate(docs_words):
counter = Counter(d)
for v in counter.keys():
_tf[v2i[v], i] = counter[v] / counter.most_common(1)[0][1]
weighted_tf = tf_methods.get(method, None)
if weighted_tf is None:
raise ValueError
return weighted_tf(_tf)
def get_idf(method="log"):
# inverse document frequency: low idf for a word appears in more docs, mean less important
df = np.zeros((len(i2v), 1))
for i in range(len(i2v)):
d_count = 0
for d in docs_words:
d_count += 1 if i2v[i] in d else 0
df[i, 0] = d_count
idf_fn = idf_methods.get(method, None)
if idf_fn is None:
raise ValueError
return idf_fn(df)
def cosine_similarity(q, _tf_idf):
unit_q = q / np.sqrt(np.sum(np.square(q), axis=0, keepdims=True))
unit_ds = _tf_idf / np.sqrt(np.sum(np.square(_tf_idf), axis=0, keepdims=True))
similarity = unit_ds.T.dot(unit_q).ravel()
return similarity
def docs_score(q, len_norm=False):
q_words = q.replace(",", "").split(" ")
# add unknown words
unknown_v = 0
for v in set(q_words):
if v not in v2i:
v2i[v] = len(v2i)
i2v[len(v2i)-1] = v
unknown_v += 1
if unknown_v > 0:
_idf = np.concatenate((idf, np.zeros((unknown_v, 1), dtype=np.float)), axis=0)
_tf_idf = np.concatenate((tf_idf, np.zeros((unknown_v, tf_idf.shape[1]), dtype=np.float)), axis=0)
else:
_idf, _tf_idf = idf, tf_idf
counter = Counter(q_words)
q_tf = np.zeros((len(_idf), 1), dtype=np.float) # [n_vocab, 1]
for v in counter.keys():
q_tf[v2i[v], 0] = counter[v]
q_vec = q_tf * _idf # [n_vocab, 1]
q_scores = cosine_similarity(q_vec, _tf_idf)
if len_norm:
len_docs = [len(d) for d in docs_words]
q_scores = q_scores / np.array(len_docs)
return q_scores
def get_keywords(n=2):
for c in range(3):
col = tf_idf[:, c]
idx = np.argsort(col)[-n:]
print("doc{}, top{} keywords {}".format(c, n, [i2v[i] for i in idx]))
tf = get_tf() # [n_vocab, n_doc]
idf = get_idf() # [n_vocab, 1]
tf_idf = tf * idf # [n_vocab, n_doc]
print("tf shape(vecb in each docs): ", tf.shape)
print("\ntf samples:\n", tf[:2])
print("\nidf shape(vecb in all docs): ", idf.shape)
print("\nidf samples:\n", idf[:2])
print("\ntf_idf shape: ", tf_idf.shape)
print("\ntf_idf sample:\n", tf_idf[:2])
# test
get_keywords()
q = "I get a coffee cup"
scores = docs_score(q)
d_ids = scores.argsort()[-3:][::-1]
print("\ntop 3 docs for '{}':\n{}".format(q, [docs[i] for i in d_ids]))
show_tfidf(tf_idf.T, [i2v[i] for i in range(tf_idf.shape[0])], "tfidf_matrix")结果：
tf shape(vecb in each docs): (47, 15)
tf samples:
[[0. 0. 0. 0. 0. 0.40546511
0.69314718 0. 0. 0. 0. 0.
0. 0. 0. ]
[0. 0.69314718 0.69314718 0. 0. 0.
0. 0. 0. 0. 0. 0.
0.40546511 0. 0. ]]
idf shape(vecb in all docs): (47, 1)
idf samples:
[[2.60943791]
[2.32175584]]
tf_idf shape: (47, 15)
tf_idf sample:
[[0. 0. 0. 0. 0. 1.05803603
1.80872453 0. 0. 0. 0. 0.
0. 0. 0. ]
[0. 1.60931851 1.60931851 0. 0. 0.
0. 0. 0. 0. 0. 0.
0.94139098 0. 0. ]]
doc0, top2 keywords ['to', 'stay']
doc1, top2 keywords ['be', 'happy']
doc2, top2 keywords ['bob', 'am']
top 3 docs for 'I get a coffee cup':
['It is coffee time, bring your cup', 'I like coffee, I like book and I like apple', 'I have a party today']
Process finished with exit code 0
1. itertools.chain(）
chain()可以把一组迭代对象串联起来，形成一个更大的迭代器：
import itertools
for c in itertools.chain('ABC', 'XYZ'):
print(c)
# 迭代效果：'A' 'B' 'C' 'X' 'Y' 'Z'2. Counter.most_common([n])
返回一个列表，其中包含 n 个最常见的元素及出现次数，按常见程度由高到低排序。 如果 n 被省略或为 None，most_common() 将返回计数器中的 所有 元素。 计数值相等的元素按首次出现的顺序排序：
Counter('abracadabra').most_common(3)
[('a', 5), ('b', 2), ('r', 2)]二、借助sklearn实现
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity
docs = [
"it is a good day, I like to stay here",
"I am happy to be here",
"I am bob",
"it is sunny today",
"I have a party today",
"it is a dog and that is a cat",
"there are dog and cat on the tree",
"I study hard this morning",
"today is a good day",
"tomorrow will be a good day",
"I like coffee, I like book and I like apple",
"I do not like it",
"I am kitty, I like bob",
"I do not care who like bob, but I like kitty",
"It is coffee time, bring your cup",
]
vectorizer = TfidfVectorizer()
tf_idf = vectorizer.fit_transform(docs)
print("idf: ", [(n, idf) for idf, n in zip(vectorizer.idf_, vectorizer.get_feature_names())])
print("v2i: ", vectorizer.vocabulary_)
q = "I get a coffee cup"
qtf_idf = vectorizer.transform([q])
res = cosine_similarity(tf_idf, qtf_idf)
res = res.ravel().argsort()[-3:]
print("\ntop 3 docs for '{}':\n{}".format(q, [docs[i] for i in res[::-1]]))
i2v = {i: v for v, i in vectorizer.vocabulary_.items()}
dense_tfidf = tf_idf.todense()结果：
idf: [('am', 2.386294361119891), ('and', 2.386294361119891), ('apple', 3.0794415416798357), ('are', 3.0794415416798357), ('be', 2.6739764335716716), ('bob', 2.386294361119891), ('book', 3.0794415416798357), ('bring', 3.0794415416798357), ('but', 3.0794415416798357), ('care', 3.0794415416798357), ('cat', 2.6739764335716716), ('coffee', 2.6739764335716716), ('cup', 3.0794415416798357), ('day', 2.386294361119891), ('do', 2.6739764335716716), ('dog', 2.6739764335716716), ('good', 2.386294361119891), ('happy', 3.0794415416798357), ('hard', 3.0794415416798357), ('have', 3.0794415416798357), ('here', 2.6739764335716716), ('is', 1.9808292530117262), ('it', 1.9808292530117262), ('kitty', 2.6739764335716716), ('like', 1.9808292530117262), ('morning', 3.0794415416798357), ('not', 2.6739764335716716), ('on', 3.0794415416798357), ('party', 3.0794415416798357), ('stay', 3.0794415416798357), ('study', 3.0794415416798357), ('sunny', 3.0794415416798357), ('that', 3.0794415416798357), ('the', 3.0794415416798357), ('there', 3.0794415416798357), ('this', 3.0794415416798357), ('time', 3.0794415416798357), ('to', 2.6739764335716716), ('today', 2.386294361119891), ('tomorrow', 3.0794415416798357), ('tree', 3.0794415416798357), ('who', 3.0794415416798357), ('will', 3.0794415416798357), ('your', 3.0794415416798357)]
v2i: {'it': 22, 'is': 21, 'good': 16, 'day': 13, 'like': 24, 'to': 37, 'stay': 29, 'here': 20, 'am': 0, 'happy': 17, 'be': 4, 'bob': 5, 'sunny': 31, 'today': 38, 'have': 19, 'party': 28, 'dog': 15, 'and': 1, 'that': 32, 'cat': 10, 'there': 34, 'are': 3, 'on': 27, 'the': 33, 'tree': 40, 'study': 30, 'hard': 18, 'this': 35, 'morning': 25, 'tomorrow': 39, 'will': 42, 'coffee': 11, 'book': 6, 'apple': 2, 'do': 14, 'not': 26, 'kitty': 23, 'care': 9, 'who': 41, 'but': 8, 'time': 36, 'bring': 7, 'your': 43, 'cup': 12}
top 3 docs for 'I get a coffee cup':
['It is coffee time, bring your cup', 'I like coffee, I like book and I like apple', 'I do not care who like bob, but I like kitty']
Process finished with exit code 0
1. numpy.ravel()
numpy的ravel() 和 flatten()函数
两者所要实现的功能是一致的（将多维数组降位一维）。这点从两个单词的意也可以看出来，ravel(散开，解开)，flatten（变平）。两者的区别在于返回拷贝（copy）还是返回视图（view），numpy.flatten()返回一份拷贝，对拷贝所做的修改不会影响（reflects）原始矩阵，而numpy.ravel()返回的是视图（view，也颇有几分C/C++引用reference的意味），会影响（reflects）原始矩阵。
In [14]: x=np.array([[1,2],[3,4]])
# flattenh函数和ravel函数在降维时默认是行序优先
In [15]: x.flatten()
Out[15]: array([1, 2, 3, 4])
In [17]: x.ravel()
Out[17]: array([1, 2, 3, 4])
# 传入'F'参数表示列序优先
In [18]: x.flatten('F')
Out[18]: array([1, 3, 2, 4])
In [19]: x.ravel('F')
Out[19]: array([1, 3, 2, 4])
#reshape函数当参数只有一个-1时表示将数组降为一维
In [21]: x.reshape(-1)
Out[21]: array([1, 2, 3, 4])
#x.T表示x的转置
In [22]: x.T.reshape(-1)
Out[22]: array([1, 3, 2, 4])2. numpy.argsort()
argsort()函数是将x中的元素从小到大排列，提取其对应的index(索引)，然后输出到y。
import numpy as np
x = np.array([1,4,3,-1,6,9])
##
y = array([3,0,2,1,4,5])
946
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
