Mining dynamic social networks from public news articles for company value prediction

Abstract

Dynamic networks are studied by sociologists to understand network evolution, belief formation, friendship formation, etc. Companies make and receive different impacts from other companies in different periods. If one can understand what types of network changes affect a company’s value, then one would be able to predict the future value of the company, grasp industry innovations, and make business more successful. However, it is often difficult to collect continuous records of network changes, and the models of mining longitudinal network are complicated. In this study, we developed algorithms and a system to infer large-scale evolutionary company networks from public news during 1981–2009. Then, based on how networks change over time, and on the financial information of the companies, we predicted company profit and revenue growth. Herein, we propose a feature extraction and selection algorithm for longitudinal networks. This paper is the first to describe a study examining longitudinal network-mining-based company performance analysis. We measured how networks affect company performance and what network features are important.