Package: shapr 1.1.0

Overview | Asymmetric conditional Shapley values | Causal Shapley values | Marginal Shapley values | Symmetric conditional Shapley values | Code example | Overview | Code setup | Symmetric conditional Shapley values (default) | Asymmetric conditional Shapley values | Symmetric marginal Shapley values | Causal Shapley values | Symmetric | Asymmetric | Comparing the frameworks | Scatter plots: marginal vs. causal Shapley values | Investigating two similar days | Sampling of coalitions | Groups of features | Implementation details | References

Vaeac | Code Examples | Basic Example | First vaeac example | Pre-trained vaeac | Pre-trained vaeac (path) | Specified max_n_coalitions | Paired sampling | Progressr | Continue the training of the vaeac approach | Vaeac with early stopping | Grouping of features | Mixed Data

The separate regression method class | Code | Setup | Linear regression model | Pre-processing | Other regression models | Cross-validation | Parallelization | The surrogate regression method class | Code | Parallelization | Add new regression methods | Summary figures | Mixed data | Mixed data: setup | Mixed data: Monte Carlo-based methods | Mixed data: separate regression methods | Mixed data: surrogate regression methods | Mixed data: summary | Regression arguments as strings | Summary | References

Introduction | Overview of Package | Functionality | Default behavior of explain | Kernel SHAP and dependence-aware estimators | The Kernel SHAP Method | Multivariate Gaussian Distribution Approach | Gaussian Copula Approach | Empirical Conditional Distribution Approach | Conditional Inference Tree Approach | Adversarial Random Forest (arf) Approach | Variational AutoEncoder with Arbitrary Conditioning (vaeac) Approach | Categorical Approach | Separate and Surrogate Regression Approaches | Estimation approaches and plotting functionality | Choosing a conditional approach | MSEv evaluation criterion | Advantage: | Disadvantages: | Confidence intervals | MSEv examples | Iterative estimation | Summary, Printing, and Result Extraction | SAGE: global feature importance | Parallelization | Batch computation | Parallelized computation | Verbosity and progress updates | Advanced usage | Combined approach | Explain groups of features | Explain custom models | Tidymodels and workflows | The parameters of the vaeac approach | Early stopping | Continued computation | Explaining a forecasting model using explain_forecast | References