Evaluating the influence of twitter on the saudi arabian stock market indicators

Abstract

Investors critically analyze past pricing history, which influences their future investment decisions. Social media and news items have a significant impact on stock market indices. In this paper, we apply machine learning and NLP principles to find the correlations between Arabic sentiments and trends in the Saudi Arabian stock market, TADAWUL. More than 277K Arabic tweets were crawled and 114K tweets were annotated manually. Three types of correlations were implemented, Pearson’s correlation coefficient, Kendall rank correlation and Spearman’s rank correlation. Moreover, the paper illustrates that the most influential users could be predictable in the future, who can have a significant impact on the stock market trends. The first achievement of this study is the collection of the largest Arabic tweets dataset specialized in finance, which will be available to the public as soon as the annotation process is finished. The second achievement is that this is the first paper to study the influence of Twitter on the Saudi stock market using different types of correlation coefficients and investigated the role of mentions on the market trends.