Automatically builds high-performance interpretable machine learning models with minimal features using a single line of code.
Auto_ViML is an automated machine learning Python library that builds multiple interpretable models with a single line of code. It automates data cleaning, feature selection, and model training to deliver high-performance models using the fewest necessary features. The tool is designed to accelerate prototyping and provide strong baseline models for classification and regression tasks.
Data scientists and machine learning practitioners who need rapid model prototyping, feature selection automation, and interpretable model generation for structured datasets.
Developers choose Auto_ViML for its emphasis on interpretability through minimal feature selection, comprehensive automation of the ML pipeline, and support for diverse data types including text and date-time variables without manual preprocessing.
Automatically Build Multiple ML Models with a Single Line of Code. Created by Ram Seshadri. Collaborators Welcome. Permission Granted upon Request.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Handles missing values, formatting, and transformations directly on raw dataframes, accepting 'dirty data' without manual preprocessing, as stated in the README.
Uses SULOV and Recursive XGBoost algorithms to automatically select the most important features, often reducing feature count by 20-99%, prioritizing interpretability.
Processes text, date-time, numeric, boolean, and categorical variables in a unified pipeline, with Auto_NLP and date-time feature generation built-in.
Integrates the imbalanced_ensemble library for skewed datasets, claiming a 5-10% boost in balanced accuracy based on the maintainer's experience.
Builds models with a single line of code, making it ideal for hackathons and baseline comparisons, as emphasized in the tips section.
The README explicitly states it's 'not meant for production,' limiting its use to prototyping and competitions, which reduces its applicability for serious deployments.
With numerous flags like Boosting_Flag, Binning_Flag, and Stacking_Flag, users may find it overwhelming to tune effectively without deep experimentation.
Automatically installs libraries like XGBoost, CatBoost, and featuretools, which can bloat environments and cause installation issues, especially in constrained setups.
Focuses on Linear, Random Forest, XGBoost, and CatBoost models, lacking support for deep learning or more exotic algorithms, which may not suit all use cases.