猫史档案馆

【作品展示】

 

【作品介绍】

LION KING

 

【作品源代码】

from nltk.classify import NaiveBayesClassifier
from nltk.corpus import movie_reviews
import nltk.classify.util
from nltk import data
from pyecharts import Liquid
import webbrowser
import os

data.path.append('./nltk_data')

#加载积极与消极评论
positive_fileids = movie_reviews.fileids('pos')
negative_fileids = movie_reviews.fileids('neg')

print(type(positive_fileids), len(positive_fileids))

#定义一个提取特征的函数
def extract_features(word_list):
    return dict([(word, True) for word in word_list])

#为积极评论贴标签
features_positive = []
for f in positive_fileids:
    word_list = movie_reviews.words(f)
    word_features = extract_features(word_list)
    features_positive.append((word_features,'Positive'))

#为消极评论贴标签
features_negative = []
for f in negative_fileids:
    word_list = movie_reviews.words(f)
    word_features = extract_features(word_list)
    features_negative.append((word_features,'Negative'))

#分成训练数据集（80%）和测试数据集（20%）
threshold_factor = 0.8
threshold_positive = int(threshold_factor * len(features_positive))
threshold_negative = int(threshold_factor * len(features_negative))

words_train = features_positive[:threshold_positive] + \
features_negative[:threshold_negative]
words_test = features_positive[threshold_positive:] + \
features_negative[threshold_negative:]
print('\n训练数据集的数量：', len(words_train))
print('\n测试数据集的数量：', len(words_test))

#训练朴素贝特斯分类器
classifier = NaiveBayesClassifier.train(words_train)
print('\n分类器的准确性：', nltk.classify.util.accuracy(classifier, words_test))

#对狮子王电影的英文评论进行预测
#提示信息
print('少年，《狮子王》电影英文评论，敢来挑战吗！')
print('游戏规则是：只需要输入英文评论，程序会判断出正负面评价，并给出预测准确度！')
print('-------------------------------------------------------------------------')

#输入评论
review = input('请在这里，输入狮王辛巴的电影评论：')
split_words = review.split()
print(split_words)
features = extract_features(split_words)
class_words = classifier.prob_classify(features)

sentiment = class_words.max()
print('预测正负面评价：', sentiment)
print('预测准确度：', round(class_words.prob(sentiment),2))

liquid = Liquid('电影《狮子王》评论：{}.\n\n正负面评价：{} \n\n预测准确度：'.format(review, sentiment))
liquid.add('Liquid', [round(class_words.prob(sentiment),2)],liquid_color=['#156ACF'])
liquid.render()
webbrowser.open('file://' + os.path.realpath('render.html'))

 

【提示】

部分含有Python第三方库相关内容的作品，在海龟编辑器网页端无法运行哦！如遇到这种情况，可以打开下面的链接，下载海龟编辑器客户端：

https://python.codemao.cn