Processing the texts in messages from social media (e.g., Twitter) to predict the expressed emotions

What is it about?

In this paper were compared different methods to predict the emotions expressed in texts coming from the social media networks (e.g., Twitter). We compared essentially two methods: the first one is based on a large set of available lexical resources that are a collection of words that experts connect to the emotions; the second one is based on machine Learning models based on latent factors. Indeed, we expected that emotions could be represented as latent factors, underlying the texts. The paper compares the two methods and discussed the differences.

Featured Image

Why is it important?

It is important to understand if emotions could be represented as latent factors underlying texts.

Perspectives