The traditional finance paradigm seeks to understand uncertainty and their impact on stock market. However, most previous studies try to quantify uncertainty at macro-level such as the EPU index. There are few studies tapping into firm-level uncertainty. In this paper, we address this empirical anomaly by integrating text mining tools to measure the firm-level uncertainty score from news content. We focus on companies listed in S&P 1500. We crawled a total of 2,196,975 news articles from LexisNexis database from April 2007 to July 2017. We extracted uncertainty related information as features by using named entity extraction, LM dictionary, and other linguistic features. We employed nonlinear machine learning models to investigate the impact on stocks future returns by uncertainty-related features. To address the theoretical problem, we use traditional asset pricing techniques to test the relationship among information derived uncertainty and the financial market performance.


Paper Number 1820; Track AI; Short Paper


When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.