We are excited to announce the pre-release of PrismaML v0.1.0! This pre-release marks the initial launch of PrismaML, a powerful library designed to streamline and simplify the machine learning model building process with comprehensive features for data analysis, feature selection, model evaluation, and visualization.
Features
Data Analysis and Visualization
- **DatasetInformation**
- `dataframe_summary`: Provides a comprehensive summary of the DataFrame, including shape, column metadata, and duplicated rows.
- `categorical_summary`: Displays a summary of categorical columns, including unique values, most frequent value, and its percentage.
- `numerical_summary`: Displays a summary of numerical columns, including statistical measures and a correlation matrix with a heatmap.
Machine Learning
- **MachineLearning**
- `select_best_features`: Evaluates and selects the best features for a given machine learning model.
- `plot_accuracy_vs_features`: Plots model performance against the number of features used.
- `evaluate_model`: Evaluates the given model over multiple train-test splits and computes average performance metrics.
- `plot_iteration_metrics`: Plots boxplots for the metrics recorded in each iteration.
Plotting
- **Plotting**
- `draw_categorical_plots`: Generates count plots for categorical columns arranged in a grid layout.
- `draw_numerical_plots`: Generates histograms for numerical columns arranged in a grid layout.
- `plot_algorithm_comparison`: Generates a bar chart comparing the performance of different algorithms based on various scores.
Installation
PrismaML is available for installation via pip and Poetry.
Contributing
We welcome contributions from the community! Whether you're interested in adding new features, fixing bugs, or improving documentation, your contributions are valuable. Please refer to our [Contribution Guide](https://github.com/Yousinator/PrismaML/blob/main/CONTRIBUTING.md) for more details on how to get started.
Discussions
Join our [Discussions](https://github.com/Yousinator/PrismaML/discussions) to share ideas, ask questions, and collaborate with other contributors. We are excited to hear your feedback and suggestions!
Code of Conduct
By participating in this project, you agree to abide by our [Code of Conduct](https://github.com/Yousinator/PrismaML/blob/main/CODEOFCONDUCT.md). We are committed to providing a welcoming and inclusive environment for everyone.
What's Next?
This pre-release is just the beginning! Here are some of the future directions we are planning:
- **DataPreprocessing**: Automate and standardize data preprocessing tasks.
- **FeatureEngineering**: Provide tools for creating new features and transforming existing features.
- **ModelBuilder**: Simplify the process of building, training, and evaluating machine learning models.
- **AutomatedML**: Automate the end-to-end machine learning workflow, from data preprocessing to model deployment.
Stay tuned for more updates and features!
Feedback
Your feedback is crucial in helping us improve PrismaML. Please feel free to open issues or discussions with your thoughts and suggestions.
Thank you for being part of the PrismaML community!
Happy coding! 💻✨