Biplots for Text and Sentiment Visualisation
Zoë-Mae Adams, Johané Nienkemper-Swanepoel
The overarching aim of this dissertation is to gain insight from text data by summarising the content and visualising the results of sentiment classification. This can be achieved by developing suitable visualisation tools for the optimal representation of sentiment classification.
The following research objectives will be investigated in pursuit of the overarching aim:
- Review of sentiment visualisation literature
- Application of adaptive sentiment lexicon to improve sentiment classification accuracy
- Enhancement of the interactive EW-MCA biplot tool
- The inspection of the ordinal nature of sentiment classification categories
- The visualisation of topic modelling results
