Analysis of Extracting Corporate Trends from News Data Using BERTopic
摘要
News is an immediate source of information. However, it is difficult to obtain information on the trends in each business unit of a company from individual news items. This study attempts to obtain information on each business unit of a company through news analysis using BERTopic by targeting multiple companies that hold similar business domains. The validity of the analysis results was verified by comparing the news analysis results with the securities reports published by the companies. As a result of the analysis, we found that (1) it is possible to follow the trends of each specific business unit in a company through news data analysis, and (2) news data contains timely information on corporate trends and product strategies, compared to securities reports. This research presents a novel analytical approach utilizing natural language processing techniques within the finance sector, highlighting a study with substantial potential for future advancements.