News-Induced Dynamic Networks for Market Signaling: Understanding the Impact of News on Firm Equity Value
提出一种文本挖掘方法，从新闻中提取企业间的共同利益/对抗利益网络，发现该网络比传统供应链和同行业网络更能预测企业股权价值，为投资策略提供新信号。
Public news provides rich information about firm operations and market dynamics. Learning about firm interactions from news is commonly done by human investors but has not been realized by automatic methods, leading to a research opportunity in market signaling via dynamic firm relations. This study proposes a new text-mining approach to extract cobenefit/counter-benefit networks based on firms’ mutual or conflicting interests in business events. It reveals that the extracted dynamic networks provide additional value in predicting firm equity value over current adopted supply chain and coindustry networks, after controlling for market activities and other traditional indicators from news, such as volume, sentiment, and comentions. In practice, such cobenefit/counter-benefit networks provide good buy and sell signals, which enrich known indicators and support more complex trading strategies in investment and portfolio management. The analysis and visualization of the extracted cobenefit/counter-benefit networks are also useful in understanding the structure of the economy and assessing firm/industry changes in a timelier fashion.