Financial Data Analysis Using Large Language Models
摘要
This paper brings new approaches on how to analyze stocks using a multi agent AI system motivated by LLMs. This system uses different AI agents that are dedicated to three tasks: research analysts, investment advisers and financial analysis. All of them combine to analyze various data including the financial status of the business, the filings made by them, marketing data and all to come up with a complete analysis for investment. A Financial Analyst Agent will review the past and present economic success of the company, the Research Analyst Agent will gauge the development of the trends by investigating the markets and the industry. The Investment Advisor Agent might utilize this information to predict future performances of stocks and give some practical advice to stakeholders. This work contains AI tools and APIs which carry out the necessary work like summarizing content from the web, looking up content on the internet as well as scrutinizing financial documents. This works with LLMs through the LangChain library as well as the YahooFinanceNewsTool and the custom agents among others. There is an additional feature that provides up to date stock prices of the company that is being analyzed.