Reading behind the tweets: A sentiment Clustering Approach

Market sentiment influence crude oil future prices in direct or indirect way. In order to measure the polarity of market sentiment various techniques has been deployed by industry and academia alike. This pilot study successfully introduced two instruments, namely topic modeling and Sentiment cluste...

Full description

Saved in:
Bibliographic Details
Published in2022 International Conference on Advanced Computing Technologies and Applications (ICACTA) pp. 1 - 6
Main Authors Saxena, Anshul, Bhagat, Vandana, Mahajan, Jayant
Format Conference Proceeding
LanguageEnglish
Published IEEE 04.03.2022
Subjects
Analytical models
Clustering algorithms
Computational modeling
Computers
Crude oil
Industries
Instruments
Oils
prices
sentiment analysis
topic modeling
tweets
Online AccessGet full text

Cover

More Information
Summary:Market sentiment influence crude oil future prices in direct or indirect way. In order to measure the polarity of market sentiment various techniques has been deployed by industry and academia alike. This pilot study successfully introduced two instruments, namely topic modeling and Sentiment clustering, to unearth the prevailing sentiments behind crude oil future pricesThree main conclusions that can be drawn from empirical results are. First, the K-Means clustering algorithm is an effective technique for sentiment clustering compared to Louveian and MDS clustering techniques. Second sentiment polarity-related positive sentiments have shown more variations in comparison to neutral and negative sentiments. Third It is possible to extract the keywords related to essential factors influencing crude oil prices using the LDA technique under topic modeling
DOI:10.1109/ICACTA54488.2022.9753107