Topic-centric sentiment analysis of UK parliamentary debates
Sentiment analysis, also known as opinion mining, is a natural language processing (NLP) task dealing with the automatic classification of text according to its conveyed sentiment. It is typically focussed on the identification of the overall sentiment---any of positive, negative or neutral---that is expressed in textual passages such as social media posts (e.g., Tweets) and online product reviews. This project is aimed at the development of approaches for a more targeted type of sentiment analysis: the identification of sentiment towards a particular subject matter, rather than just detecting overall sentiment (i.e., regardless of topic), from political discourse.
As part of this work, the PhD student will investigate and develop methods for: (1) topic detection, in which the topics being discussed in a given piece of text are identified; and (2) topic-centric sentiment analysis, which involves the analysis of sentiments specific to the automatically detected topics. The methods will be applied to the analysis of documents in the UK Hansard archives, which contain historical textual records of parliamentary debates and written statements.