Wind Power Forecasting Application

Overview

This project involves the development and deployment of a wind power forecasting application leveraging machine learning and deep learning techniques. The application predicts wind power using key variables such as wind speed, wind direction, and theoretical power. A user-friendly web interface was built using Streamlit for real-time predictions.

---

Features

• Machine Learning Models:

  • Implemented models like Linear Regression, XGBoost, and Ensemble methods.
  • Achieved R² = 0.9667 with effective hyperparameter tuning and cross-validation.

• Deep Learning Models:

  • Developed and optimized neural network architectures:
    • Single Hidden Layer.
    • Multiple Hidden Layers.
    • Long Short-Term Memory (LSTM) networks.
  • Best LSTM model achieved R² = 0.92.

• Time Series Analysis:

  • Incorporated temporal dependencies to improve prediction accuracy.

• Feature Engineering:

  • Conducted advanced preprocessing, including:
    • Multivariate signal decomposition.
    • Variable selection.

• Web Application:

  • Designed a user-friendly interface using Streamlit.
  • Integrated machine learning models for real-time wind power forecasts.

---

Technology Stack

• Programming Languages: Python
• Machine Learning Libraries: scikit-learn, XGBoost
• Deep Learning Libraries: TensorFlow/Keras, PyTorch
• Web Framework: Streamlit

---

Results

• Machine Learning Performance:
  • Achieved high prediction accuracy with Linear Regression and XGBoost (R² = 0.9667).
• Deep Learning Performance:
  • Best LSTM model demonstrated robust predictive capabilities (R² = 0.92).

---

How to Run

1. Clone the Repository:

   git clone https://github.com/username/wind-power-forecasting.git
   cd wind-power-forecasting

2. Install Dependencies:

   pip install -r requirements.txt

3. Run the Streamlit Application:

   streamlit run app.py

---

Dataset

The dataset contains key variables for wind power forecasting, such as:

• Wind Speed
• Wind Direction
• Theoretical Power

The dataset was preprocessed to handle missing values, outliers

---

Skills Highlighted

• Machine Learning (Linear Regression, XGBoost, Ensemble Methods)
• Deep Learning (Neural Networks, LSTM)
• Time Series Analysis
• Streamlit Web Application Development
• Feature Engineering

---