Tweet Classification of Topic and Emotion
Twitter is a key communication tool.
The vast amount of opinions exchanged via Twitter can provide insight into not only events or topics of importance for the users,
but also into the mood of users expressing their opinions on Twitter.
Correlating the topics discussed on Tweeter with the mood of the Tweeter users, as conveyed by their writing, may aid with identifying potential areas of concern.
For example, a tweet on a political topic conveying angry sentiment, may indicate that the topic of discussion may require monitoring.
This project aims to perform topic classification and emotion modeling on Twitter data set(s).
We are looking to classify tweets into the following topics: politics, religion, and family.
Then we are looking to classify the mood of the tweets into the following categories: happiness, depression, and anger.
Last, we will correlate the topic with the mood to determine what topics should be monitored further.
Proposed data set --
see Stanford link under Training Data.
- Review of the literature on this topic
- Review of the literature on machine learning classification techniques
- Design, create, and test a classification system to classify the data set into the following topic categories: politics, religion, family
- Design, create, and test a classification system to classify the mood of the Twitter user into the following categories: happiness, depression, anger
- Correlate each topic with each mood and report your findings
Bollen, Johan, Huina Mao, and Alberto Pepe.
Modeling public mood and emotion: Twitter sentiment and socio-economic phenomena,
ICWSM 11 (2011): 450-453.
Pak, Alexander, and Patrick Paroubek.
Twitter as a Corpus for Sentiment Analysis and Opinion Mining,
LREc. Vol. 10. No. 2010. 2010.