Predicting Saudi Stock Market Index by Incorporating GDELT Using Multivariate Time Series Modelling

Prediction of financial and economic markets is very challenging but valuable for economists, business owners, and traders. Forecasting stock market prices depends on many factors, such as other markets’ performance, economic state of a country, and others. In behavioral finance, people’s emotions and opinions influence their transactional decisions and therefore the financial markets. The focus of this research is to predict the Saudi Stock Market Index by utilizing its previous values and the impact of people’s sentiments on their financial decisions. Human emotions and opinions are directly influenced by media and news, which we incorporated by utilizing the Global Data on Events, Location, and Tone (GDELT) dataset by Google. GDELT is a collection of news from all over the world from different types of media such as TV, broadcasts, radio, newspapers, and websites. We extracted two time series from GDELT, filtered for Saudi Arabian news. The two time series represent daily values of tone and social media attention. We studied the characteristics of the generated multivariate time series, then deployed and compared multiple multivariate models to predict the daily index of the Saudi stock market.