AI Stock Price Prediction Using Large Language Models in Python

This tutorial demonstrates how to build a combined AI and machine learning pipeline to predict stock prices with just a laptop. I use Meta Llama 3, an advanced open source large language model (LLM), with a local Ollama server on my laptop to run a sentiment analysis of recent financial news headlines scraped from the web using a Python API library. This sentiment data is then used to create a measure to predict stock prices using time series forecasting methods. The analysis is run through an interactive Streamlit web app, making it user friendly.

The goal of this tutorial is to show you the process of how to use AI transformer neural network models to gain insight into stock prices with a Python code workflow. A more comprehensive analysis would include additional time series training data for stock prices and news headlines, a deep dive into the accuracy of the Llama 3 model on classifying sentiment, the addition of other relevant features for predicting stock price movement, and a more deliberate decision for the ML modeling method.

****Important Note: This video is not financial or investing advice. It is an educational tutorial on how to use AI models within a machine learning pipeline.***

Like, Comment, and Subscribe to the Deep Charts Channel for more informative AI and Machine Learning tutorials.

**Data Science Resources**

[BitCuration]

@DeepCharts, do you have your code in github or a way to download? Would like to prototype based on this video. Nice works.

[RodMorrison]

Great video, love the art of the possible!

[osmanniazi]

Not a bad intro at all. I am an ex Goldman Sachs Quant. I dont know how youtube got me here :-). But I think this is good for someone new to Quant finance and machine learning. Yes someone needs to think deeply about the pricing but this is a good starting point to know how to use these tools.

[NoneOfTheAboveEntParty]

@DeepCharts - I love the idea... I'm still new to python and learning how to do things. When I follow your instructions it is giving me an error allowed to merge between different levels. (1 levels on the left, 2 on the right)" as if the dataframes are not aligned. any suggestions on how to resolve this?

[jonathanmitchell]

Appreciate an AI finance video that focuses on handling/ presenting data as information rather than placing trades. I work a salary job with family and i just dont have the time to proper DD and sometimes my subscriptions go unused for a month or so. I am looking to integrate local AI into my strategy by helping make sense of web articles, reports, and analysts sentiments/ratings. Presenting this data in a manageable format in real time. I am on AMD system so Pytorch makes sense? If use this as template to learn on am I on the right track?

[brooklyn_domino]

How are you deploying the app on Streamlit? I keep getting this error: "ConnectionError: HTTPConnectionPool(host='0.0.0.0', port=11434): Max retries

exceeded with url". I've tried changing the server address from 127.0.0.1 to 0.0.0.0. I've restarted the server, changed the base_url but nothing seems to work. I'm missing some trick here. Do I need to make sure that the ollama server is running on the local system when executing the web app? Because for some reason when I run the program locally from vscode it works.

[uwaishkarni]

I think without model training and RL technique it wouldn't be accurate

[Charles-mj]

Yeah so that is not how the stock market works.