LTC Stocks - Time Series Prediction with Liquid Neural Networks
Motivation:
Exploring the potential of liquid neural networks for time series prediction, specifically in forecasting stock prices. Liquid neural networks adapt their internal structure dynamically, making them well-suited for complex data like financial markets.
Work:
- Implemented a stock price prediction model using liquid neural networks.
- Collected and processed historical stock price data for training.
- Evaluated model performance against traditional architectures like LSTMs and GRUs.
- Conducted experiments to test adaptability and prediction accuracy over volatile market periods.
Why It Matters:
Stock price prediction is a challenging time series problem due to market volatility. Liquid neural networks offer a promising approach by dynamically adjusting to new data, potentially outperforming static models in uncertain environments.