This Streamlit app demonstrates how machine learning works, focusing on simple classification algorithms for predicting wine quality and colour. Users can experiment with data, adjust parameters, and select different models to observe how these changes affect the model's performance.
Predict wine quality or colour using machine learning models. Choose from different models: Random Forest, XGBoost, or Decision Tree. Adjust model parameters and hyperparameters. Visualise model performance metrics and feature importance. Interactive user input for wine characteristics. Display of ROC curves for model evaluation.
Select what to predict: Quality or Colour, Choose a machine learning model, Adjust test size and model hyperparameters, Input custom wine characteristics or use default values, View model performance metrics, feature importance, and ROC curves.
The app uses the UCI Machine Learning Repository wine dataset, which includes various chemical components of wines along with their quality ratings and colours. Model Performance Metrics The app displays the following performance metrics: Accuracy, Precision, Recall, F1-score.
Top 5 Most Important Features bar chart, ROC Curve (binary for Colour prediction, multi-class for Quality prediction).
Python, Streamlit, Pandas, Scikit-learn, XGBoost, Matplotlib.
To run the app locally: Clone this repository, Install the required packages: pip install -r requirements.txt, Run the Streamlit app: streamlit run streamlit_app.py.
Streamlit Documentation, Scikit-learn Documentation, XGBoost Documentation, UCI Wine Quality Dataset.
This app was created by Bloch AI LTD. It is distributed under the GNU GENERAL PUBLIC LICENSE, Version 3. For more information, visit www.bloch.ai.