Skip to content

Latest commit

 

History

History
57 lines (36 loc) · 2.53 KB

readme.md

File metadata and controls

57 lines (36 loc) · 2.53 KB

Identifying Outliers in Well Logs using Machine Learning Models

Welcome to the GitHub repository for my Streamlit app designed to identify outliers in well logs! This application is an essential tool for geoscientists and engineers who want to clean and analyze well log data efficiently.

🔗 Access the App and Manual

  • Access the App: To start using the app, click here.
  • App Manual: For a detailed explanation of how to use this app, click here. Note that the manual is available in Spanish.

🤖 Machine Learning Models

The app utilizes the following machine learning models for outlier detection:

  • Isolation Forest
  • OneClass SVM (Support Vector Machine)
  • LOF (Local Outlier Factor)

🛠️ Interact with Hyperparameters

Users can interact with the hyperparameters of each machine learning model to customize the outlier detection process according to their specific needs. Explore the options available in the app to fine-tune the models and optimize outlier detection results.

🚀 Features

  • Data Upload: Easily upload your well log data.
  • Outlier Detection: Automatically or manually detect and visualize outliers in your data.
  • Data Cleaning: Options to remove or adjust detected outliers.
  • Visualization: Interactive charts, boxplots and scatterplot to understand your data better.

💡 How to Use

  1. Access the App: Open the Streamlit app.
  2. Upload Data: Load your well log files.
  3. Analyze: Use the tools provided to detect and manage outliers in your dataset.
  4. Visualize Outliers: View visualizations highlighting the detected outliers in your well log data.

📚 Resources

For further information and resources, refer to the following:

🤝 Contributing

Contributions are welcome! If you have suggestions for improving the app, please feel free to fork this repository, make changes, and submit a pull request.

📞 Contact

If you have any questions or would like to discuss the app further, please contact me via.