News Text Analysis: Industry Co-occurrence Changes
factor.formula
Industry co-occurrence change factor calculation formula:
in:
- :
The number of industries that appear in news texts at the same time as stock i within the time interval t (e.g., one trading day, one week).
- :
The number of industries that appear in news texts at the same time as stock i in the t-1 time interval.
- :
The month-on-month change in the number of industry co-occurrences of stock i in period t relative to period t-1 is the industry co-occurrence change factor. This factor can be understood as the increase or decrease in the industry relevance of stock i.
factor.explanation
This factor is based on news text analysis. It captures the potential information diffusion effect and changes in industry correlation in the market by calculating the month-on-month changes in the number of industries that appear in news reports with a specific stock. When a stock frequently appears in the news with multiple industries, it may imply that the company's business connections, supply chain relationships, or market sentiment interactions with these industries are strengthening. In the A-share market, it is observed that this factor is negatively correlated with future returns, which may reflect that when a stock appears in the news with more industries, it is often accompanied by higher market attention, and high attention may not be conducive to excess returns of stock prices in the short term. Conversely, when a stock appears in the news with fewer industries, it may mean that the market's attention to it has decreased, but there may be opportunities for excess returns. Therefore, this factor can be used as a reverse indicator for short-term return forecasts, while reflecting investors' perception and expected changes in stock information flow.