Are there any training data of tweets with labels: positive, negative, neutral as following:
"tweet 1" positive
"tweet 2" positive
"tweet 3" neutral
"tweet 4" negative
Or in general are there any good site for finding training sets?
>Solution :
You can find good training datasets on kaggle!
A good example would be the Sentiment140 dataset with over 1.6 million tweets. They have been annotated with (0 = negative, 4 = positive) – but it’s explained on-site! 🙂