Honours Thesis Defence - TEXT MINING ON TWITTER USING LDA TOPIC MODEL AND SENTIMENT ANALYSIS
HONOURS THESIS DEFENCE
Jodrey School of Computer Science
TEXT MINING ON TWITTER USING LDA TOPIC MODEL AND SENTIMENT ANALYSIS
Friday, February 9, 2018 at 1:30 PM
Carnegie Hall - Room 207
Twitter is one of the most popular microblogging platforms, where millions of users daily share their attitudes, views, and opinions. Thus, Twitter has become one of the richest research sources for text mining. The primary goal of this research is to explore text mining using two approaches: LDA topic modelling and sentiment analysis by examining Twitter plain English text data.
Latent Dirichlet Allocation (LDA) topic model is a probabilistic model which can find the hidden topics in the document. LDA topic model examines the document as a mixture of topics, where each topic is combined with related words with certain probabilities. The results are clear and easier to understand. This method makes it easy to analyze a large set of tweets and conclude the main topics using a set of words. Sentiment analysis is a method to extract information from the opinions, attitudes, and feelings of individuals.
During the presentation, the background knowledge of LDA topic model and sentiment analysis will be explained. The experiment design, preliminary work and implement process will be discussed. The results of the experiment will reveal that the proposed techniques are effective and significant in practical use.
Supervisor: Haiyi Zhang
Second Reader: André Trudel