xploring Domain Specific Sentiment from Stock Investor Tweets

  • D. Deepa, S.Selvaraj et al.


In recent years the financial industry keeps on boost up its analytics capabilities to leverage
significantly higher actionable insights from user feedback in order to increase return on investment.
Owing to the raise of public participation in social media, large volumes of beliefs are expressed
which are again unstructured in nature. Determine the future value of the stock market is expected to
be robust, accurate and efficient. The stock prediction is to be done in the way by correlating the
public sentiment with the behavior of stock prices and by bearing in mind all the variables that might
affect the stock value and performance. Natural language processing provides best feature extraction
techniques and Machine learning offer better classification algorithms jointly provides powerful
results. This paper performs an analysis on various feature extraction techniques to detect the
polarity of words from the stocktwits then classify the opinion using Logistic Regression classifier.
Evaluation is done with feature engineering techniques like CountVectorizer, TF-IDF, Word2Vec and
Glove by using machine learning.

How to Cite
et al., D. D. S. (2020). xploring Domain Specific Sentiment from Stock Investor Tweets. International Journal of Advanced Science and Technology, 29(6s), 2067-2073. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/10915