Team-068

Inspiration

Our project, inspired by "The Wolf of Wall Street," aims to leverage sentiment analysis from social media feeds to make informed decisions in stock trading.

What it does

The application calculates a probability distribution for sentiment analysis from incoming social media data. Based on the analysis outcomes, strategic decisions are made to buy or sell stocks.

How we built it

Our model is constructed using the pre-trained Language Model (LLM) from Distilbert due to its outstanding performance, providing us with a competitive advantage. The training dataset received from Optiver was carefully transformed to be labeled as negative, neutral, or positive. The predictions are then integrated into a Python bot script for executing stock transactions.

Challenges we ran into

Several challenges tested our team's resilience, including the competitiveness of dual stock trading, determining the optimal neural network architecture, selecting an effective trading strategy, collaborative coding in a shared virtual machine, and achieving a positive Profit and Loss (PnL).

Accomplishments that we're proud of

We take pride in achieving a positive PnL, the efficiency and accuracy of our sentiment analysis model, successful teamwork, and the creation of an engaging video for the project. (We hope you enjoy)

What we learned

Throughout the project, we gained valuable insights into financial markets, deepened our understanding of neural networks, especially in the context of transfer learning, and acquired essential skills for participating in hackathons.