Media-based Corporate Network and Its Effects on Stock Market.

We study the comoving patterns of relevant stocks in terms of media co-exposure in both normal and extreme market periods. For this purpose, we build a media-based corporate network in terms of 17,685 pieces of news articles released in 2014 that mentioned at least two stocks listed on the CSI 300 Index. Each node in the network represents a listed company. The edge weight between two nodes is determined by the number of news articles that mention these two firms. We find that the comoving patterns, which were discovered by the media co-exposure, of relevant stocks are stable in both normal and crisis periods. The stock performance of one listed firm is affected by its neighboring firms in the proposed media-based corporate network. These results are consistent with the theoretical models of noise and liquidity traders. [ABSTRACT FROM AUTHOR]