SENTIMENT AND MARKET ANALYSIS OF ENERGY SECTOR USING NATURAL LANGUAGE PROCESSING AND MACHINE LEARNING TECHNIQUES
Aryan Singh, Mainak Ghosh, Kumaran M., A. Sharmila
School of Electrical Engineering,Vellore Institute of Technology, VIT University,
Vellore – 632014, Tamil Nadu, India
ABSTRACT: Crude Oil is an essential natural resource that has a big impact on the world economy, and sentiment analysis or opinion mining can often be tricky, especially when it comes to detection of emotions from news excerpts when they may be diplomatically phrased- we as humans tend to express ourselves differently and one sentence alone can have detections of different polarities. The focus of this paper is to propose an oil price trend prediction method by inferring sentiments from online news and social media. This is done by comparing the working and accuracies of the three most popular models used for opinion mining- VADER, TextBlob and Multinomial Naive-Bayes. Our aim is to analyze market sentiments using text recognition by comparing the accuracies of the two NLP-based models- VADER and TextBlob, and the Naive-Bayes model based on machine learning. While sentiment alone cannot always predict changes in commodity prices, with the help of technical analysis tools, better insights can be gained to determine in this case, the spot prices of Crude Oil.
INDEX TERMS: Naive Bayes, VADER, Textblob, NLP, Crude Oil, Sentiment Analysis.