将word文本文件导入python，文件地址为C:\Users\Admin\Desktop\三国演义.docx，用jieba实现词频统计，输出前10个频率最高的词，并构建20个节点的知识图谱

首先需要安装jieba和wordcloud两个库：在命令行中输入

pip install jieba
pip install wordcloud

然后可以按以下步骤实现：

1. 导入库

import jieba
import wordcloud
from docx import Document

2. 读取文本

doc = Document('C:\\Users\\Admin\\Desktop\\三国演义.docx')
text = ''
for p in doc.paragraphs:
    text += p.text

3. 分词

words = jieba.lcut(text)

4. 统计词频

freq = {}
for word in words:
    if len(word) == 1:
        continue
    freq[word] = freq.get(word, 0) + 1

5. 输出前10个频率最高的词

top10 = sorted(freq.items(), key=lambda x: x[1], reverse=True)[:10]
for word, count in top10:
    print(word, count)

6. 构建知识图谱

import networkx as nx
import matplotlib.pyplot as plt

G = nx.Graph()
for i, (word1, count1) in enumerate(freq.items()):
    if i >= 20:
        break
    G.add_node(word1)
    for j, (word2, count2) in enumerate(freq.items()):
        if i >= j:
            continue
        if word2.startswith(word1) or word1.startswith(word2):
            weight = min(count1, count2)
            G.add_edge(word1, word2, weight=weight)

pos = nx.spring_layout(G, k=0.5)
nx.draw(G, pos, with_labels=True, font_size=10, node_color='y', node_size=1000, edge_color='gray', width=1.5, alpha=0.8)
nx.draw_networkx_edge_labels(G, pos, edge_labels={(u, v): G[u][v]['weight'] for u, v in G.edges()}, font_size=8)
plt.show()

完整代码如下：

import jieba
import wordcloud
from docx import Document
import networkx as nx
import matplotlib.pyplot as plt

# 读取文本
doc = Document('C:\\Users\\Admin\\Desktop\\三国演义.docx')
text = ''
for p in doc.paragraphs:
    text += p.text

# 分词
words = jieba.lcut(text)

# 统计词频
freq = {}
for word in words:
    if len(word) == 1:
        continue
    freq[word] = freq.get(word, 0) + 1

# 输出前10个频率最高的词
top10 = sorted(freq.items(), key=lambda x: x[1], reverse=True)[:10]
for word, count in top10:
    print(word, count)

# 构建知识图谱
G = nx.Graph()
for i, (word1, count1) in enumerate(freq.items()):
    if i >= 20:
        break
    G.add_node(word1)
    for j, (word2, count2) in enumerate(freq.items()):
        if i >= j:
            continue
        if word2.startswith(word1) or word1.startswith(word2):
            weight = min(count1, count2)
            G.add_edge(word1, word2, weight=weight)

pos = nx.spring_layout(G, k=0.5)
nx.draw(G, pos, with_labels=True, font_size=10, node_color='y', node_size=1000, edge_color='gray', width=1.5, alpha=0.8)
nx.draw_networkx_edge_labels(G, pos, edge_labels={(u, v): G[u][v]['weight'] for u, v in G.edges()}, font_size=8)
plt.show()