Stock Price Forecasting
LSTM-based time series model for stock price predictions with Flask API
Project Overview
This project leverages Long Short-Term Memory (LSTM) neural networks to forecast stock prices based on historical data. The model analyzes time series data to predict future trends, achieving an 18% improvement in RMSE over baseline models. It is deployed via a Flask API for real-time predictions, providing actionable insights for traders and investors.
Technologies Used
- Keras/TensorFlow: For building and training the LSTM model.
- Python: Core language for data processing and modeling.
- Pandas & NumPy: For data manipulation and analysis.
- Matplotlib: For visualizing predictions and trends.
- Flask: For deploying the model as a real-time API.
Key Features
- Accurate stock price predictions using LSTM networks.
- Preprocessing pipeline for handling financial datasets (e.g., Yahoo Finance).
- Interactive visualizations of predictions vs. actual prices.
- Hyperparameter tuning for optimized performance (18% RMSE reduction).
- Real-time predictions via a scalable Flask API.
Model Evaluation Metrics
- RMSE: 0.12 (18% improvement over baseline).
- MAE: 0.09.
- R² Score: 0.85.
Impact
The model empowers traders and investors with reliable predictions, reducing errors by 18% compared to traditional methods. The Flask API enables seamless integration into trading platforms, enhancing decision-making in volatile markets.