Sentiment analysis of social media and its relation to stock market
|Author:||Bc. Vladimír Zobal,|
|Year:||2017 - summer|
|Leaders:|| prof. PhDr. Ladislav Krištoufek Ph.D.
|Work type:|| Bachelors
|Awards and prizes:|
|Abstract:||The aim of this thesis is to verify and quantify relationship between the sentiment
of Twitter posts, so called Tweets, and shot-run movements of the
stock market indices. Firstly, the sentiment analysis, using lexicon-based
approach, is conducted and subsequently predictions of stock market indices
for the next day are made. Our analysis focuses mainly on noise reduction
by implementation of threshold, and also on differences between negative
and positive sentiment in contrast with reaction of individual market actors.
All predictions are conducted on daily basis and the main purpose of this
model is to support trading decision of individual by inclusion of sentiment
score obtained from Twitter.