挖掘词云

# -*- coding: utf-8 -*-

# 代码12-1 评论去重的代码

import pandas as pd

import re

import jieba.posseg as psg

import numpy as np

# 去重，去除完全重复的数据

reviews = pd.read_csv("../../data/0404/reviews.csv")

reviews = reviews[['content', 'content_type']].drop_duplicates()

content = reviews['content']

# 代码12-2 数据清洗

# 去除去除英文、数字等

# 由于评论主要为京东美的电热水器的评论，因此去除这些词语

strinfo = re.compile('[0-9a-zA-Z]|京东|美的|电热水器|热水器|')

content = content.apply(lambda x: strinfo.sub('', x))

# 代码12-3 分词、词性标注、去除停用词代码

# 分词

worker = lambda s: [(x.word, x.flag) for x in psg.cut(s)] # 自定义简单分词函数

seg_word = content.apply(worker)

# 将词语转为数据框形式，一列是词，一列是词语所在的句子ID，最后一列是词语在该句子的位置

n_word = seg_word.apply(lambda x: len(x))  # 每一评论中词的个数

n_content = [[x+1]*y for x,y in zip(list(seg_word.index), list(n_word))]

index_content = sum(n_content, [])  # 将嵌套的列表展开，作为词所在评论的id

seg_word = sum(seg_word, [])

word = [x[0] for x in seg_word]  # 词

nature = [x[1] for x in seg_word]  # 词性

content_type = [[x]*y for x,y in zip(list(reviews['content_type']), list(n_word))]

content_type = sum(content_type, [])  # 评论类型

result = pd.DataFrame({"index_content":index_content,

                       "word":word,

                       "nature":nature,

                       "content_type":content_type})

# 删除标点符号

result = result[result['nature'] != 'x']  # x表示标点符号

# 删除停用词

stop_path = open("../../data/0404/stoplist.txt", 'r',encoding='UTF-8')

stop = stop_path.readlines()

stop = [x.replace('\n', '') for x in stop]

word = list(set(word) - set(stop))

result = result[result['word'].isin(word)]

# 构造各词在对应评论的位置列

n_word = list(result.groupby(by = ['index_content'])['index_content'].count())

index_word = [list(np.arange(0, y)) for y in n_word]

index_word = sum(index_word, [])  # 表示词语在改评论的位置

# 合并评论id，评论中词的id，词，词性，评论类型

result['index_word'] = index_word

# 代码12-4 提取含有名词的评论

# 提取含有名词类的评论

ind = result[['n' in x for x in result['nature']]]['index_content'].unique()

result = result[[x in ind for x in result['index_content']]]

# 代码12-5 绘制词云

import matplotlib.pyplot as plt

from wordcloud import WordCloud

frequencies = result.groupby(by = ['word'])['word'].count()

frequencies = frequencies.sort_values(ascending = False)

# print(frequencies)

backgroud_Image=plt.imread('../../data/0404/pl.jpg')

wordcloud = WordCloud(font_path="C:/Windows/Fonts/STZHONGS.ttf",

                      max_words=100,

                      background_color='white',

                      mask=backgroud_Image)

my_wordcloud = wordcloud.fit_words(frequencies)

plt.title('3106',size=20)

plt.imshow(my_wordcloud)

plt.axis('off')

plt.show()

# 将结果写出

result.to_csv("../../data/0404/word.csv", index = False, encoding = 'utf-8')

 

 

# -*- coding: utf-8 -*-

# 代码12-6 匹配情感词

import pandas as pd

import numpy as np

word = pd.read_csv("../../data/0404/word.csv")

# 读入正面、负面情感评价词

pos_comment = pd.read_csv("../../data/0404/正面评价词语（中文）.txt", header=None,sep="\n",

                          encoding = 'utf-8', engine='python')

neg_comment = pd.read_csv("../../data/0404/负面评价词语（中文）.txt", header=None,sep="\n",

                          encoding = 'utf-8', engine='python')

pos_emotion = pd.read_csv("../../data/0404/正面情感词语（中文）.txt", header=None,sep="\n",

                          encoding = 'utf-8', engine='python')

neg_emotion = pd.read_csv("../../data/0404/负面情感词语（中文）.txt", header=None,sep="\n",

                          encoding = 'utf-8', engine='python')

# 合并情感词与评价词

positive = set(pos_comment.iloc[:,0])|set(pos_emotion.iloc[:,0])

negative = set(neg_comment.iloc[:,0])|set(neg_emotion.iloc[:,0])

intersection = positive&negative  # 正负面情感词表中相同的词语

positive = list(positive - intersection)

negative = list(negative - intersection)

positive = pd.DataFrame({"word":positive,

                         "weight":[1]*len(positive)})

negative = pd.DataFrame({"word":negative,

                         "weight":[-1]*len(negative)})

posneg = positive.append(negative)

#  将分词结果与正负面情感词表合并，定位情感词

data_posneg = posneg.merge(word, left_on = 'word', right_on = 'word',

                           how = 'right')

data_posneg = data_posneg.sort_values(by = ['index_content','index_word'])

# 代码12-7 修正情感倾向

# 根据情感词前时候有否定词或双层否定词对情感值进行修正

# 载入否定词表

notdict = pd.read_csv("../../data/0404/not.csv")

# 处理否定修饰词

data_posneg['amend_weight'] = data_posneg['weight']  # 构造新列，作为经过否定词修正后的情感值

data_posneg['id'] = np.arange(0, len(data_posneg))

only_inclination = data_posneg.dropna()  # 只保留有情感值的词语

only_inclination.index = np.arange(0, len(only_inclination))

index = only_inclination['id']

for i in np.arange(0, len(only_inclination)):

    review = data_posneg[data_posneg['index_content'] ==

                         only_inclination['index_content'][i]]  # 提取第i个情感词所在的评论

    review.index = np.arange(0, len(review))

    affective = only_inclination['index_word'][i]  # 第i个情感值在该文档的位置

    if affective == 1:

        ne = sum([i in notdict['term'] for i in review['word'][affective - 1]])

        if ne == 1:

            data_posneg['amend_weight'][index[i]] = -\

            data_posneg['weight'][index[i]]          

    elif affective > 1:

        ne = sum([i in notdict['term'] for i in review['word'][[affective - 1,

                  affective - 2]]])

        if ne == 1:

            data_posneg['amend_weight'][index[i]] = -\

            data_posneg['weight'][index[i]]

           

# 更新只保留情感值的数据

only_inclination = only_inclination.dropna()

# 计算每条评论的情感值

emotional_value = only_inclination.groupby(['index_content'],

                                           as_index=False)['amend_weight'].sum()

# 去除情感值为0的评论

emotional_value = emotional_value[emotional_value['amend_weight'] != 0]

# 代码12-8 查看情感分析效果

# 给情感值大于0的赋予评论类型（content_type）为pos,小于0的为neg

emotional_value['a_type'] = ''

emotional_value['a_type'][emotional_value['amend_weight'] > 0] = 'pos'

emotional_value['a_type'][emotional_value['amend_weight'] < 0] = 'neg'

# 查看情感分析结果

result = emotional_value.merge(word,

                               left_on = 'index_content',

                               right_on = 'index_content',

                               how = 'left')

result = result[['index_content','content_type', 'a_type']].drop_duplicates()

confusion_matrix = pd.crosstab(result['content_type'], result['a_type'],

                               margins=True)  # 制作交叉表

(confusion_matrix.iat[0,0] + confusion_matrix.iat[1,1])/confusion_matrix.iat[2,2]

# 提取正负面评论信息

ind_pos = list(emotional_value[emotional_value['a_type'] == 'pos']['index_content'])

ind_neg = list(emotional_value[emotional_value['a_type'] == 'neg']['index_content'])

posdata = word[[i in ind_pos for i in word['index_content']]]

negdata = word[[i in ind_neg for i in word['index_content']]]

# 绘制词云

import matplotlib.pyplot as plt

from wordcloud import WordCloud

# 正面情感词词云

freq_pos = posdata.groupby(by = ['word'])['word'].count()

freq_pos = freq_pos.sort_values(ascending = False)

backgroud_Image=plt.imread('../../data/0404/pl.jpg')

wordcloud = WordCloud(font_path="C:/Windows/Fonts/STZHONGS.ttf",

                      max_words=100,

                      background_color='white',

                      mask=backgroud_Image)

pos_wordcloud = wordcloud.fit_words(freq_pos)

plt.imshow(pos_wordcloud)

plt.title('3106',size=20)

plt.axis('off')

plt.show()

# 负面情感词词云

freq_neg = negdata.groupby(by = ['word'])['word'].count()

freq_neg = freq_neg.sort_values(ascending = False)

neg_wordcloud = wordcloud.fit_words(freq_neg)

plt.imshow(neg_wordcloud)

plt.title('3106',size=20)

plt.axis('off')

plt.show()

# 将结果写出,每条评论作为一行

posdata.to_csv("../../data/0404/posdata.csv", index = False, encoding = 'utf-8')

negdata.to_csv("../../data/0404/negdata.csv", index = False, encoding = 'utf-8')

 

 

 

 

# -*- coding: utf-8 -*-

# 代码12-9 建立词典及语料库

import pandas as pd

import numpy as np

import re

import itertools

import matplotlib.pyplot as plt

# 载入情感分析后的数据

posdata = pd.read_csv("../../data/0404/posdata.csv", encoding = 'utf-8')

negdata = pd.read_csv("../../data/0404/negdata.csv", encoding = 'utf-8')

from gensim import corpora, models

# 建立词典

pos_dict = corpora.Dictionary([[i] for i in posdata['word']])  # 正面

neg_dict = corpora.Dictionary([[i] for i in negdata['word']])  # 负面

# 建立语料库

pos_corpus = [pos_dict.doc2bow(j) for j in [[i] for i in posdata['word']]]  # 正面

neg_corpus = [neg_dict.doc2bow(j) for j in [[i] for i in negdata['word']]]   # 负面

# 代码12-10 主题数寻优

# 构造主题数寻优函数

def cos(vector1, vector2):  # 余弦相似度函数

    dot_product = 0.0;  

    normA = 0.0;  

    normB = 0.0;  

    for a,b in zip(vector1, vector2):

        dot_product += a*b  

        normA += a**2  

        normB += b**2  

    if normA == 0.0 or normB==0.0:  

        return(None)  

    else:  

        return(dot_product / ((normA*normB)**0.5))  

# 主题数寻优

def lda_k(x_corpus, x_dict):  

   

    # 初始化平均余弦相似度

    mean_similarity = []

    mean_similarity.append(1)

   

    # 循环生成主题并计算主题间相似度

    for i in np.arange(2,11):

        lda = models.LdaModel(x_corpus, num_topics = i, id2word = x_dict)  # LDA模型训练

        for j in np.arange(i):

            term = lda.show_topics(num_words = 50)

           

        # 提取各主题词

        top_word = []

        for k in np.arange(i):

            top_word.append([''.join(re.findall('"(.*)"',i)) \

                             for i in term[k][1].split('+')])  # 列出所有词

           

        # 构造词频向量

        word = sum(top_word,[])  # 列出所有的词  

        unique_word = set(word)  # 去除重复的词

       

        # 构造主题词列表，行表示主题号，列表示各主题词

        mat = []

        for j in np.arange(i):

            top_w = top_word[j]

            mat.append(tuple([top_w.count(k) for k in unique_word]))  

           

        p = list(itertools.permutations(list(np.arange(i)),2))

        l = len(p)

        top_similarity = [0]

        for w in np.arange(l):

            vector1 = mat[p[w][0]]

            vector2 = mat[p[w][1]]

            top_similarity.append(cos(vector1, vector2))

           

        # 计算平均余弦相似度

        mean_similarity.append(sum(top_similarity)/l)

    return(mean_similarity)

           

# 计算主题平均余弦相似度

pos_k = lda_k(pos_corpus, pos_dict)

neg_k = lda_k(neg_corpus, neg_dict)        

# 绘制主题平均余弦相似度图形

from matplotlib.font_manager import FontProperties  

font = FontProperties(size=14)

#解决中文显示问题

plt.rcParams['font.sans-serif']=['SimHei']

plt.rcParams['axes.unicode_minus'] = False  

fig = plt.figure(figsize=(10,8))

ax1 = fig.add_subplot(211)

ax1.plot(pos_k)

ax1.set_xlabel('3106', fontproperties=font)

ax2 = fig.add_subplot(212)

ax2.plot(neg_k)

ax2.set_xlabel('3106', fontproperties=font)

# 代码12-11 LDA主题分析

# LDA主题分析

pos_lda = models.LdaModel(po

 

 

s_corpus, num_topics = 3, id2word = pos_dict)  

neg_lda = models.LdaModel(neg_corpus, num_topics = 3, id2word = neg_dict)  

pos_lda.print_topics(num_words = 10)

neg_lda.print_topics(num_words = 10)