利用人工智能挖掘财报会议纪要中的投资与风险管理机遇

Core Viewpoint - The article discusses the innovative approach of using large language models (LLMs) to analyze earnings call transcripts, enabling analysts to assess the sentiment of CEOs regarding future business outlooks and their potential impact on stock prices [1][2]. Group 1: Advanced Earnings Call Analysis - LSEG MarketPsych Transcript Analytics integrates LSEG's data resources with MarketPsych's natural language processing (NLP) capabilities, providing sentiment analysis and thematic data for over 16,000 publicly listed companies [2][3]. - The solution identifies over 1,000 themes and 4,000 event types within earnings call transcripts, allowing for detailed sentiment classification and analysis [3][4]. Group 2: Application Scenarios - Companies with high sentiment scores in earnings calls tend to outperform those with lower scores in the following month, indicating a correlation between CEO sentiment and stock performance [6]. - The built-in ESG sentiment classifier can dynamically monitor ESG-related sentiments, providing risk warnings for companies with low ESG sentiment scores [6][7]. - The analysis system can also quantify the frequency and sentiment of key negative terms mentioned by executives, aiding in risk management and credit risk monitoring [7].