With the growth of artificial intelligence, new opportunities have opened up to analyze and predict real-time data. This project uses large language models (LLMs) to predict stock market trends. LLMs are advanced AI tools designed to understand and generate text, which helps finding patterns in large, complex datasets. The goal is to use real-time stock market data, including prices, trading volumes, and historical trends, to train the LLM to recognize patterns and predict short-term changes in the market with high accuracy.

The project will follow a complete process to achieve this. First, stock market data is collected using tools such as YFinance that gives the access to the financial data on yahoo finance.The python based libraries in Yfinance converted the financial data into a spreadsheet indicating the company's stock information. It included the open and closed stock values along with the high and low values during that day. This time series stock data for 25 companies was gathered and various algorithms were used to develop the predictive models for their stocks. Initial data was modeled by using the auto ARIMA algorithm. Day by day acquired data was stored in a Retrieval-Augmented Generation (RAG) database.

This database efficiently manages updates to make sure the model always works with the latest data. The acquired data is then cleaned and formatted for the LLMs (ChatGPT, LLMA etc), emphasizing patterns and trends essential for prediction. Both are employed and are being evaluated for prediction accuracy, speed, and adaptability to rapid market changes. This comparison will provide details into the strengths and weaknesses of each model for financial forecasting. These combined technologies can build a clear workflow of using AI to analyze stock market trends, address challenges in a dynamic financial environment, and to improve these tools in the future.