A Streamlit-powered web app that forecasts stock prices using advanced time series models: ARIMA, SARIMA, Prophet, and LSTM. Built to compare model performance using RMSE and MAPE metrics.
📊 Features
- 🧠 Supports 4 forecasting models:
- ARIMA
- SARIMA
- Prophet
- LSTM (Keras)
- 📅 Select forecast horizon (7 / 30 / 90 days)
- 📉 Visual forecast plots with prediction lines
- 📦 Performance comparison (RMSE / MAPE)
- 📁 Data loading and cleaning utilities
- 🖼️ Saves forecast plots as PNG
🛠️ Tech Stack
- Python 3.10
- Streamlit
- pandas
- matplotlib
- Prophet
- statsmodels
- TensorFlow / Keras
🗂️ Project Structure
stock-forecasting-app/ ├── data/ │ └── raw_data.csv ├── models/ │ ├── arima_model.py │ ├── sarima_model.py │ ├── prophet_model.py │ └── lstm_model.py ├── utils/ │ └── preprocessing.py ├── outputs/ │ └── forecast images ├── streamlit_app.py ├── get_data.py ├── metrics.py └── README.md
🚀 Getting Started
- Clone the Repository
git clone https://github.com/abhijit1620/stock-forecasting-app.git cd stock-forecasting-app
- Run App
streamlit run streamlit_app.py
** Author **
Abhijeet Sharma BTech CSE-AIML | Lucknow [email protected]
Chandana R MCA| Bengaluru [email protected]