如何确定某事的 'did' 或 'did not'

Question

区分这两者的直接方法是什么：

1. the movie received critical acclaim
2. the movie did not attain critical acclaim.

在我看来 'sentiment analysis' 的 nlp 可以为我做这件事。所以我使用 Textblob sentiment analysis。但是两个句子的polarity都是0.0.

Answer 1

它需要否定处理能力。例如，wink-nlp supports negation handling. You can checkout the code with this example at runkit.

Answer 2

对于simple，您可以使用基于LSTM的flair模型，将单词序列考虑在内进行预测。

1.安装 flair

!pip3 install flair

2。代码

import flair
flair_sentiment = flair.models.TextClassifier.load('en-sentiment')

sentence1 = 'the movie received critical acclaim'
sentence2 = 'the movie did not attain critical acclaim'

s1 = flair.data.Sentence(sentence1)
flair_sentiment.predict(s1)
s1_sentiment = s1.labels
print(s1_sentiment)

s2 = flair.data.Sentence(sentence2)
flair_sentiment.predict(s2)
s2_sentiment = s2.labels
print(s2_sentiment)

3。结果

print(s1_sentiment)
[POSITIVE (0.9995)]

print(s2_sentiment)
[NEGATIVE (0.9985)]

更多关于flair的详情，您可以访问this github repo.