Predicting Daily Price Action for the Oil and Gas Industry with Machine Learning Technology using Twitter Sentiment
Open AccessIn the oil and gas industry, Engineering managers are critical in directing capital projects and ensuring they align with market sentiment and financial forecasts. This is a complex task because the oil and gas industry is volatile in nature. The pandemic highlighted the effect of social media on investment decisions, it also exposed predatory short selling strategies that can sometimes impede project expansion. However, companies that report strong goodwill can benefit from positive sentiment, which may lead to increased daily stock prices. In this domain, social media has emerged as a complex factor impacting these dynamics. This Praxis developed Long Short-Term Memory (LSTM) daily predictive price models as tools to examine the impact of Twitter sentiment and short interest on daily stock prices for Occidental Petroleum Corporation (OXY). Multiple predictors were used in the LSTM models including OXY Open (OXYopen), Manually applied Twitter sentiment and sentiment deriving from Linear Support Vector Classifiers (LSVC) and Bidirectional Encoder Representations from Transformers (BERT), OXY short interest (OXYsi), US strategic petroleum reserves (SPR), trading price per barrel of oil (CLFopen), US global exchange rate (USD), and average US gas prices (AUSGP). This Praxis shows that contrary to the literature, Twitter sentiment and OXY short interest were not significant predictors of daily OXY Open stock prices.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.