How fmartinssp/malta works
MALTA is a research tool for exploring and treating missing data. Its documented capabilities cover data loading and visualization, generating MCAR/MAR/MNAR omissions, testing missingness mechanisms, imputing values, classifying outcomes, and evaluating imputations; a protocol summarizes selected treatment methods. The README names these capabilities but does not describe runtime calls or data transfers, and no source excerpts were provided, so the graph shows components without inferring workflow edges.
Components 12
Data exploration
- Load data
src/load_data/load_data.ipynb - Missing-data visualization
src/visualization/missing_data_visualization.ipynb - Generate missingness
src/missing_data_generation/missing_values_generator.ipynb
Missingness analysis
- Little test
src/missing_mechanism_detection/little_test.ipynb - Dixon test
src/missing_mechanism_detection/dixon_test.ipynb - Homoscedasticity test
src/missing_mechanism_detection/homoscedasticity_test.ipynb - Regression-based test
src/missing_mechanism_detection/rb_test.ipynb
Imputation and prediction
- Imputation methods
src/imputation/imputation_methods.ipynb - Classifiers
src/classifiers/classifiers.ipynb
Evaluation and protocol
- Evaluation metrics
src/evaluation/metrics.ipynb - Handling protocol
src/protocol/Missing_Data_Handling_Protocol.ipynb
Other components
- Researcher