Date of Award
January 2016
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Economics & Finance
First Advisor
Xiao Wang
Abstract
In this paper, I attempt to apply an emotional proxy derived by applying
the Affective Norms for English Words (ANEW) to messages posted to the
Twitter social networking service in order to forecast the movement two stock
market indices: the Dow Jones Industrial Average (DJIA) and the CBOE
Volatility Index (VIX). In contrast to previous works, I have compared the
results of various forecast models employing different sentiment variables, as
well as comparing the neural network approach to more standard logistic re-
gression. Additionally, several of the models used employ an as-yet unique
sentiment proxy, focusing on the average of expressed emotion rather than the
volume of expressed emotion. The results indicate that while there is a distinct
possibility that sentiment variables can assist in accurately forecasting market
movement, the differences in choice of sentiment proxy and forecast method
are less important than anticipated.
Recommended Citation
Plenzick, Christopher, "Can You Really Predict Markets With Twitter?" (2016). Theses and Dissertations. 2064.
https://commons.und.edu/theses/2064