Sentiment Analysis on Financial News and Microblogs
Abstract
Sentiment analysis is useful for multiple tasks including customer satisfaction metrics, identifying market trends for any industry or products, analyzing reviews from social media comments. This thesis highlights the importance of sentiment analysis, provides a summary of seminal works and different approaches towards sentiment analysis. It aims to address sentiment analysis on financial news and microblogs by classifying textual data from financial news and microblogs as positive or negative. Sentiment analysis is performed by making use of paragraph vectors and logistic regression in this thesis and it aims to compare it with previously performed approaches to performing analysis and help researchers in this field. This approach achieves state of the art results for the dataset used in this research. It also presents an insightful analysis of the results of this approach.
Degree
M.S.
Advisors
Rayz, Purdue University.
Subject Area
Linguistics|Information Technology|Artificial intelligence
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.