这是对涂铭等老师撰写的《Python自然语言处理实战:核心技术与算法》中第9章NLP中用到的机器学习算法
的学习笔记。
文本分类:中文垃圾邮件分类
- 特征提取器构建(前期准备)
"""
@author: liushuchun
"""
from sklearn.feature_extraction.text import CountVectorizer
def bow_extractor(corpus, ngram_range=(1, 1)):
vectorizer = CountVectorizer(min_df=1, ngram_range=ngram_range)
features = vectorizer.fit_transform(corpus)
return vectorizer, features
from sklearn.feature_extraction.text import TfidfTransformer
def tfidf_transformer(bow_matrix):
transformer = TfidfTransformer(norm='l2',
smooth_idf=True,
use_idf=True)
tfidf_matrix = transformer.fit_transform(bow_matrix)
return transformer, tfidf_matrix
from sklearn.feature_extraction.text import TfidfVectorizer
def tfidf_extractor(corpus, ngram_range=(1, 1)):
vectorizer = TfidfVectorizer(min_df=1,
norm='l2',
smooth_idf=True,
use_idf=True,
ngram_range=ngram_range)
features = vectorizer.fit_transform(corpus)
return vectorizer, features
- 数据标注处理(前期准备)
"""
@author: liushuchun
"""
import re
import string
import jieba
# 加载停用词
with open("dict/stop_words.utf8", encoding="utf8") as f:
stopword_list = f.readlines()
def tokenize_text(text):
tokens = jieba.cut(text)
tokens = [token.strip() for token in tokens]
return tokens
def remove_special_characters(text):
tokens = tokenize_text(text)
pattern = re.compile('[{}]'.format(re.escape(string.punctuation)))
filtered_tokens = filter(None, [pattern.sub('', token) for token in tokens])
filtered_text = ' '.join(filtered_tokens)
return filtered_text
def remove_stopwords(text):
tokens = tokenize_text(text)
filtered_tokens = [token for token in tokens if token not in stopword_list]
filtered_text = ''.join(filtered_tokens)
return filtered_text
def normalize_corpus(corpus, tokenize=False):
normalized_corpus = []
for text in corpus:
text = remove_special_characters(text)
text = remove_stopwords(text)
normalized_corpus.append(text)
if tokenize:
text = tokenize_text(text)
normalized_corpus.append(text)
return normalized_corpus
(3)邮件分类全流程(只需运行这个,把前两个文件放在同一路径下)
"""
author: liushuchun
"""
import numpy as np
from sklearn.model_selection import train_test_split
def get_data():
'''
获取数据
:return: 文本数据,对应的labels
'''
with open("data/ham_data.txt", encoding="utf8") as ham_f, open("data/spam_data.txt", encoding="utf8") as spam_f:
ham_data = ham_f.readlines()
spam_data = spam_f.readlines()
ham_label = np.ones(len(ham_data)).tolist()
spam_label = np.zeros(len(spam_data)).tolist()
corpus = ham_data + spam_data
labels = ham_label + spam_label
return corpus, labels
def prepare_datasets(corpus, labels, test_data_proportion=0.3):
'''
:param corpus: 文本数据
:param labels: label数据
:param test_data_proportion:测试数据占比
:return: 训练数据,测试数据,训练label,测试label
'''
train_X, test_X, train_Y, test_Y = train_test_split(corpus, labels,
test_size=test_data_proportion, random_state=42)
return train_X, test_X, train_Y, test_Y
def remove_empty_docs(corpus, labels):
filtered_corpus = []
filtered_labels = []
for doc, label in zip(corpus, labels):
if doc.strip():
filtered_corpus.append(doc)
filtered_labels.append(label)
return filtered_corpus, filtered_labels
from sklearn import metrics
def get_metrics(true_labels, predicted_labels):
print('准确率:', np.round(
metrics.accuracy_score(true_labels,
predicted_labels),
2))
print('精度:', np.round(
metrics.precision_score(true_labels,
predicted_labels,
average='weighted'),
2))
print('召回率:', np.round(
metrics.recall_score(true_labels,
predicted_labels,
average='weighted'),
2))
print('F1得分:', np.round(
metrics.f1_score(true_labels,
predicted_labels,
average='weighted'),
2))
def train_predict_evaluate_model(classifier,
train_features, train_labels,
test_features, test_labels):
# build model
classifier.fit(train_features, train_labels)
# predict using model
predictions = classifier.predict(test_features)
# evaluate model prediction performance
get_metrics(true_labels=test_labels,
predicted_labels=predictions)
return predictions
def main():
corpus, labels = get_data() # 获取数据集
print("总的数据量:", len(labels))
corpus, labels = remove_empty_docs(corpus, labels)
print('样本之一:', corpus[10])
print('样本的label:', labels[10])
label_name_map = ["垃圾邮件", "正常邮件"]
print('实际类型:', label_name_map[int(labels[10])], label_name_map[int(labels[5900])])
# 对数据进行划分
train_corpus, test_corpus, train_labels, test_labels = prepare_datasets(corpus,
labels,
test_data_proportion=0.3)
from normalization import normalize_corpus
# 进行归一化
norm_train_corpus = normalize_corpus(train_corpus)
norm_test_corpus = normalize_corpus(test_corpus)
''.strip()
from feature_extractors import bow_extractor, tfidf_extractor
import gensim
import jieba
# 词袋模型特征
bow_vectorizer, bow_train_features = bow_extractor(norm_train_corpus)
bow_test_features = bow_vectorizer.transform(norm_test_corpus)
# tfidf 特征
tfidf_vectorizer, tfidf_train_features = tfidf_extractor(norm_train_corpus)
tfidf_test_features = tfidf_vectorizer.transform(norm_test_corpus)
# tokenize documents
tokenized_train = [jieba.lcut(text)
for text in norm_train_corpus]
print(tokenized_train[2:10])
tokenized_test = [jieba.lcut(text)
for text in norm_test_corpus]
# build word2vec 模型
model = gensim.models.Word2Vec(tokenized_train,
size=500,
window=100,
min_count=30,
sample=1e-3)
from sklearn.naive_bayes import MultinomialNB
from sklearn.linear_model import SGDClassifier
from sklearn.linear_model import LogisticRegression
mnb = MultinomialNB()
svm = SGDClassifier(loss='hinge', n_iter_no_change=100)
lr = LogisticRegression()
# 基于词袋模型的多项朴素贝叶斯
print("基于词袋模型特征的贝叶斯分类器")
mnb_bow_predictions = train_predict_evaluate_model(classifier=mnb,
train_features=bow_train_features,
train_labels=train_labels,
test_features=bow_test_features,
test_labels=test_labels)
# 基于词袋模型特征的逻辑回归
print("基于词袋模型特征的逻辑回归")
lr_bow_predictions = train_predict_evaluate_model(classifier=lr,
train_features=bow_train_features,
train_labels=train_labels,
test_features=bow_test_features,
test_labels=test_labels)
# 基于词袋模型的支持向量机方法
print("基于词袋模型的支持向量机")
svm_bow_predictions = train_predict_evaluate_model(classifier=svm,
train_features=bow_train_features,
train_labels=train_labels,
test_features=bow_test_features,
test_labels=test_labels)
# 基于tfidf的多项式朴素贝叶斯模型
print("基于tfidf的贝叶斯模型")
mnb_tfidf_predictions = train_predict_evaluate_model(classifier=mnb,
train_features=tfidf_train_features,
train_labels=train_labels,
test_features=tfidf_test_features,
test_labels=test_labels)
# 基于tfidf的逻辑回归模型
print("基于tfidf的逻辑回归模型")
lr_tfidf_predictions=train_predict_evaluate_model(classifier=lr,
train_features=tfidf_train_features,
train_labels=train_labels,
test_features=tfidf_test_features,
test_labels=test_labels)
# 基于tfidf的支持向量机模型
print("基于tfidf的支持向量机模型")
svm_tfidf_predictions = train_predict_evaluate_model(classifier=svm,
train_features=tfidf_train_features,
train_labels=train_labels,
test_features=tfidf_test_features,
test_labels=test_labels)
import re
num = 0
for document, label, predicted_label in zip(test_corpus, test_labels, svm_tfidf_predictions):
if label == 0 and predicted_label == 0:
print('邮件类型:', label_name_map[int(label)])
print('预测的邮件类型:', label_name_map[int(predicted_label)])
print('文本:-')
print(re.sub('\n', ' ', document))
num += 1
if num == 4:
break
num = 0
for document, label, predicted_label in zip(test_corpus, test_labels, svm_tfidf_predictions):
if label == 1 and predicted_label == 0:
print('邮件类型:', label_name_map[int(label)])
print('预测的邮件类型:', label_name_map[int(predicted_label)])
print('文本:-')
print(re.sub('\n', ' ', document))
num += 1
if num == 4:
break
if __name__ == "__main__":
main()
文本聚类实战:用K-means 对豆瓣读书数据聚类
- 爬取豆瓣读书数据
import ssl
import bs4
import re
import requests
import csv
import codecs
import time
from urllib import request, error
context = ssl._create_unverified_context()
class DouBanSpider:
def __init__(self):
self.userAgent = "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.115 Safari/537.36"
self.headers = {"User-Agent": self.userAgent}
# 拿到豆瓣图书的分类标签
def getBookCategroies(self):
try:
url = "https://book.douban.com/tag/?view=type&icn=index-sorttags-all"
response = request.urlopen(url, context=context)
content = response.read().decode("utf-8")
return content
except error.HTTPError as identifier:
print("errorCode: " + identifier.code + "errrorReason: " + identifier.reason)
return None
# 找到每个标签的内容
def getCategroiesContent(self):
content = self.getBookCategroies()
if not content:
print("页面抓取失败...")
return None
soup = bs4.BeautifulSoup(content, "lxml")
categroyMatch = re.compile(r"^/tag/*")
categroies = []
for categroy in soup.find_all("a", {"href": categroyMatch}):
if categroy:
categroies.append(categroy.string)
return categroies
# 拿到每个标签的链接
def getCategroyLink(self):
categroies = self.getCategroiesContent()
categroyLinks = []
for item in categroies:
link = "https://book.douban.com/tag/" + str(item)
categroyLinks.append(link)
return categroyLinks
def getBookInfo(self, categroyLinks):
self.setCsvTitle()
categroies = categroyLinks
try:
for link in categroies:
print("正在爬取:" + link)
bookList = []
response = requests.get(link)
soup = bs4.BeautifulSoup(response.text, 'lxml')
bookCategroy = soup.h1.string
for book in soup.find_all("li", {"class": "subject-item"}):
bookSoup = bs4.BeautifulSoup(str(book), "lxml")
bookTitle = bookSoup.h2.a["title"]
bookAuthor = bookSoup.find("div", {"class": "pub"})
bookComment = bookSoup.find("span", {"class": "pl"})
bookContent = bookSoup.li.p
# print(bookContent)
if bookTitle and bookAuthor and bookComment and bookContent:
bookList.append([bookTitle.strip(),bookCategroy.strip() , bookAuthor.string.strip(),
bookComment.string.strip(), bookContent.string.strip()])
self.saveBookInfo(bookList)
time.sleep(3)
print("爬取结束....")
except error.HTTPError as identifier:
print("errorCode: " + identifier.code + "errrorReason: " + identifier.reason)
return None
def setCsvTitle(self):
csvFile = codecs.open("data/data.csv", 'a', 'utf_8_sig')
try:
writer = csv.writer(csvFile)
writer.writerow(['title', 'tag', 'info', 'comments', 'content'])
finally:
csvFile.close()
def saveBookInfo(self, bookList):
bookList = bookList
csvFile = codecs.open("data/data.csv", 'a', 'utf_8_sig')
try:
writer = csv.writer(csvFile)
for book in bookList:
writer.writerow(book)
finally:
csvFile.close()
def start(self):
categroyLink = self.getCategroyLink()
self.getBookInfo(categroyLink)
douBanSpider = DouBanSpider()
douBanSpider.start()
- 数据标准化处理
"""
@author: liushuchun
"""
import re
import string
import jieba
# 加载停用词
with open("dict/stop_words.utf8", encoding="utf8") as f:
stopword_list = f.readlines()
def tokenize_text(text):
tokens = jieba.lcut(text)
tokens = [token.strip() for token in tokens]
return tokens
def remove_special_characters(text):
tokens = tokenize_text(text)
pattern = re.compile('[{}]'.format(re.escape(string.punctuation)))
filtered_tokens = filter(None, [pattern.sub('', token) for token in tokens])
filtered_text = ' '.join(filtered_tokens)
return filtered_text
def remove_stopwords(text):
tokens = tokenize_text(text)
filtered_tokens = [token for token in tokens if token not in stopword_list]
filtered_text = ''.join(filtered_tokens)
return filtered_text
def normalize_corpus(corpus):
normalized_corpus = []
for text in corpus:
text =" ".join(jieba.lcut(text))
normalized_corpus.append(text)
return normalized_corpus
- 文本聚类全流程(只需运行这个,把前两个文件放在同一路径下)
```python
"""
@author: liushuchun
"""
import pandas as pd
import numpy as np
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
def build_feature_matrix(documents, feature_type='frequency',
ngram_range=(1, 1), min_df=0.0, max_df=1.0):
feature_type = feature_type.lower().strip()
if feature_type == 'binary':
vectorizer = CountVectorizer(binary=True,
max_df=max_df, ngram_range=ngram_range)
elif feature_type == 'frequency':
vectorizer = CountVectorizer(binary=False, min_df=min_df,
max_df=max_df, ngram_range=ngram_range)
elif feature_type == 'tfidf':
vectorizer = TfidfVectorizer()
else:
raise Exception("Wrong feature type entered. Possible values: 'binary', 'frequency', 'tfidf'")
feature_matrix = vectorizer.fit_transform(documents).astype(float)
return vectorizer, feature_matrix
book_data = pd.read_csv('data/data.csv') #读取文件
print(book_data.head())
book_titles = book_data['title'].tolist()
book_content = book_data['content'].tolist()
print('书名:', book_titles[0])
print('内容:', book_content[0][:10])
from normalization import normalize_corpus
# normalize corpus
norm_book_content = normalize_corpus(book_content)
# 提取 tf-idf 特征
vectorizer, feature_matrix = build_feature_matrix(norm_book_content,
feature_type='tfidf',
min_df=0.2, max_df=0.90,
ngram_range=(1, 2))
# 查看特征数量
print(feature_matrix.shape)
# 获取特征名字
feature_names = vectorizer.get_feature_names()
# 打印某些特征
print(feature_names[:10])
from sklearn.cluster import KMeans
def k_means(feature_matrix, num_clusters=10):
km = KMeans(n_clusters=num_clusters,
max_iter=10000)
km.fit(feature_matrix)
clusters = km.labels_
return km, clusters
num_clusters = 10
km_obj, clusters = k_means(feature_matrix=feature_matrix,
num_clusters=num_clusters)
book_data['Cluster'] = clusters
from collections import Counter
# 获取每个cluster的数量
c = Counter(clusters)
print(c.items())
def get_cluster_data(clustering_obj, book_data,
feature_names, num_clusters,
topn_features=10):
cluster_details = {}
# 获取cluster的center
ordered_centroids = clustering_obj.cluster_centers_.argsort()[:, ::-1]
# 获取每个cluster的关键特征
# 获取每个cluster的书
for cluster_num in range(num_clusters):
cluster_details[cluster_num] = {}
cluster_details[cluster_num]['cluster_num'] = cluster_num
key_features = [feature_names[index]
for index
in ordered_centroids[cluster_num, :topn_features]]
cluster_details[cluster_num]['key_features'] = key_features
books = book_data[book_data['Cluster'] == cluster_num]['title'].values.tolist()
cluster_details[cluster_num]['books'] = books
return cluster_details
def print_cluster_data(cluster_data):
# print cluster details
for cluster_num, cluster_details in cluster_data.items():
print('Cluster {} details:'.format(cluster_num))
print('-' * 20)
print('Key features:', cluster_details['key_features'])
print('book in this cluster:')
print(', '.join(cluster_details['books']))
print('=' * 40)
import matplotlib.pyplot as plt
from sklearn.manifold import MDS
from sklearn.metrics.pairwise import cosine_similarity
import random
from matplotlib.font_manager import FontProperties
def plot_clusters(num_clusters, feature_matrix,
cluster_data, book_data,
plot_size=(16, 8)):
# generate random color for clusters
def generate_random_color():
color = '#%06x' % random.randint(0, 0xFFFFFF)
return color
# define markers for clusters
markers = ['o', 'v', '^', '<', '>', '8', 's', 'p', '*', 'h', 'H', 'D', 'd']
# build cosine distance matrix
cosine_distance = 1 - cosine_similarity(feature_matrix)
# dimensionality reduction using MDS
mds = MDS(n_components=2, dissimilarity="precomputed",
random_state=1)
# get coordinates of clusters in new low-dimensional space
plot_positions = mds.fit_transform(cosine_distance)
x_pos, y_pos = plot_positions[:, 0], plot_positions[:, 1]
# build cluster plotting data
cluster_color_map = {}
cluster_name_map = {}
for cluster_num, cluster_details in cluster_data[0:500].items():
# assign cluster features to unique label
cluster_color_map[cluster_num] = generate_random_color()
cluster_name_map[cluster_num] = ', '.join(cluster_details['key_features'][:5]).strip()
# map each unique cluster label with its coordinates and books
cluster_plot_frame = pd.DataFrame({'x': x_pos,
'y': y_pos,
'label': book_data['Cluster'].values.tolist(),
'title': book_data['title'].values.tolist()
})
grouped_plot_frame = cluster_plot_frame.groupby('label')
# set plot figure size and axes
fig, ax = plt.subplots(figsize=plot_size)
ax.margins(0.05)
# plot each cluster using co-ordinates and book titles
for cluster_num, cluster_frame in grouped_plot_frame:
marker = markers[cluster_num] if cluster_num < len(markers) \
else np.random.choice(markers, size=1)[0]
ax.plot(cluster_frame['x'], cluster_frame['y'],
marker=marker, linestyle='', ms=12,
label=cluster_name_map[cluster_num],
color=cluster_color_map[cluster_num], mec='none')
ax.set_aspect('auto')
ax.tick_params(axis='x', which='both', bottom='off', top='off',
labelbottom='off')
ax.tick_params(axis='y', which='both', left='off', top='off',
labelleft='off')
fontP = FontProperties()
fontP.set_size('small')
ax.legend(loc='upper center', bbox_to_anchor=(0.5, -0.01), fancybox=True,
shadow=True, ncol=5, numpoints=1, prop=fontP)
# add labels as the film titles
for index in range(len(cluster_plot_frame)):
ax.text(cluster_plot_frame.ix[index]['x'],
cluster_plot_frame.ix[index]['y'],
cluster_plot_frame.ix[index]['title'], size=8)
# show the plot
plt.show()
cluster_data = get_cluster_data(clustering_obj=km_obj,
book_data=book_data,
feature_names=feature_names,
num_clusters=num_clusters,
topn_features=5)
print_cluster_data(cluster_data)
plot_clusters(num_clusters=num_clusters,
feature_matrix=feature_matrix,
cluster_data=cluster_data,
book_data=book_data,
plot_size=(16, 8))
from sklearn.cluster import AffinityPropagation
def affinity_propagation(feature_matrix):
sim = feature_matrix * feature_matrix.T
sim = sim.todense()
ap = AffinityPropagation()
ap.fit(sim)
clusters = ap.labels_
return ap, clusters
# get clusters using affinity propagation
ap_obj, clusters = affinity_propagation(feature_matrix=feature_matrix)
book_data['Cluster'] = clusters
# get the total number of books per cluster
c = Counter(clusters)
print(c.items())
# get total clusters
total_clusters = len(c)
print('Total Clusters:', total_clusters)
cluster_data = get_cluster_data(clustering_obj=ap_obj,
book_data=book_data,
feature_names=feature_names,
num_clusters=total_clusters,
topn_features=5)
print_cluster_data(cluster_data)
plot_clusters(num_clusters=num_clusters,
feature_matrix=feature_matrix,
cluster_data=cluster_data,
book_data=book_data,
plot_size=(16, 8))
from scipy.cluster.hierarchy import ward, dendrogram
def ward_hierarchical_clustering(feature_matrix):
cosine_distance = 1 - cosine_similarity(feature_matrix)
linkage_matrix = ward(cosine_distance)
return linkage_matrix
def plot_hierarchical_clusters(linkage_matrix, book_data, figure_size=(8, 12)):
# set size
fig, ax = plt.subplots(figsize=figure_size)
book_titles = book_data['title'].values.tolist()
# plot dendrogram
ax = dendrogram(linkage_matrix, orientation="left", labels=book_titles)
plt.tick_params(axis='x',
which='both',
bottom='off',
top='off',
labelbottom='off')
plt.tight_layout()
plt.savefig('ward_hierachical_clusters.png', dpi=200)
# build ward's linkage matrix
linkage_matrix = ward_hierarchical_clustering(feature_matrix)
# plot the dendrogram
plot_hierarchical_clusters(linkage_matrix=linkage_matrix,
book_data=book_data,
figure_size=(8, 10))