Sentiment Correlation in News Network and Associated Market Movements Based on Co-occurrence Network

The rapid development of media allows anyone to disseminate information quickly and efficiently. In an increasingly connected global market, the sentiment of news on companies not only shows the performance of the market itself but also influences the movement of the broader market. Through this information, investors make an appropriate investment decision to maximize profits. Therefore, all investment-related information is needed to assist investors in making decisions. This work, we apply NLP techniques to learn the news sentiment of 22 company based on CNN news for 6 years. We conduct the research to show the sentiment correlation to the market movements based on the news co-occurrence network. The financial market is a system with complex inter-dependencies between companies. The co-occurrence of companies in the news may reveal the underlying relations between the companies. Sentiment analysis annals each news's positive, neutral, and negative classes and evaluates the company's sentiment based on their co-occurrence in the same news article. We find groups of highly related companies generally align with the sectoral classification. The network uncovers more interesting information when the groups and sectors different, potentially containing insights above sector classification offers.