简介
VADER（Valence-Aware Dictionary and mootion Reasoner）是一个词汇和基于规则的情感分析工具。
论文
名称：VADER: A Parsimonious Rule-Based Model for Sentiment Analysis of Social Media Text
论文地址:https://www.aaai.org/ocs/index.php/ICWSM/ICWSM14/paper/viewPaper/8109
发表于AAAI 2014
安装
地址：https://pypi.org/project/vaderSentiment/
Github源码：https://github.com/cjhutto/vaderSentiment
安装方法: pip install vaderSentiment
使用
使用方法
from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer
sentences = ["This is awful","I like it"]
analyzer = SentimentIntensityAnalyzer()
for sentence in sentences:
vs = analyzer.polarity_scores(sentence)
print("{:-<65} {}".format(sentence, str(vs)))
输出
This is awful---------------- {'neg': 0.6, 'neu': 0.4, 'pos': 0.0, 'compound': -0.4588}
I like it-------------------- {'neg': 0.0, 'neu': 0.286, 'pos': 0.714, 'compound': 0.3612}
存储形式是字典，通过字典调用各个变量
解释
- neg 负面情绪
- neu 中立情绪
- pos 积极情绪
- compound 指的是将前三个指标相加，并调整为(-1,1)的一个值