R-universe search: wmshap

Showing 2 of total 2 results (show query)

haghish

HMDA:Holistic Multimodel Domain Analysis for Exploratory Machine Learning

Holistic Multimodel Domain Analysis (HMDA) is a robust and transparent framework designed for exploratory machine learning research, aiming to enhance the process of feature assessment and selection. HMDA addresses key limitations of traditional machine learning methods by evaluating the consistency across multiple high-performing models within a fine-tuned modeling grid, thereby improving the interpretability and reliability of feature importance assessments. Specifically, it computes Weighted Mean SHapley Additive exPlanations (WMSHAP), which aggregate feature contributions from multiple models based on weighted performance metrics. HMDA also provides confidence intervals to demonstrate the stability of these feature importance estimates. This framework is particularly beneficial for analyzing complex, multidimensional datasets common in health research, supporting reliable exploration of mental health outcomes such as suicidal ideation, suicide attempts, and other psychological conditions. Additionally, HMDA includes automated procedures for feature selection based on WMSHAP ratios and performs dimension reduction analyses to identify underlying structures among features. For more details see Haghish (2025) <doi:10.13140/RG.2.2.32473.63846>.

Maintained by E. F. Haghish. Last updated 17 hours ago.

ensemble-feature-importance explainable-ai explainable-artificial-intelligence explainable-machine-learning explainable-ml exploratory-machine-learning exploratory-modelling feature-importance feature-selection-methods holistic-modeling holistic-multimodel-domain-analysis multimodel-ensemble reproducible-ai reproducible-research robust-feature-selection shapley-additive-explanations shapley-values transparent-ai weighted-mean-shap wmshap

16.0 match 1 stars 3.54 score

haghish

shapley:Weighted Mean SHAP and CI for Robust Feature Assessment in ML Grid

This R package introduces Weighted Mean SHapley Additive exPlanations (WMSHAP), an innovative method for calculating SHAP values for a grid of fine-tuned base-learner machine learning models as well as stacked ensembles, a method not previously available due to the common reliance on single best-performing models. By integrating the weighted mean SHAP values from individual base-learners comprising the ensemble or individual base-learners in a tuning grid search, the package weights SHAP contributions according to each model's performance, assessed by multiple either R squared (for both regression and classification models). alternatively, this software also offers weighting SHAP values based on the area under the precision-recall curve (AUCPR), the area under the curve (AUC), and F2 measures for binary classifiers. It further extends this framework to implement weighted confidence intervals for weighted mean SHAP values, offering a more comprehensive and robust feature importance evaluation over a grid of machine learning models, instead of solely computing SHAP values for the best model. This methodology is particularly beneficial for addressing the severe class imbalance (class rarity) problem by providing a transparent, generalized measure of feature importance that mitigates the risk of reporting SHAP values for an overfitted or biased model and maintains robustness under severe class imbalance, where there is no universal criteria of identifying the absolute best model. Furthermore, the package implements hypothesis testing to ascertain the statistical significance of SHAP values for individual features, as well as comparative significance testing of SHAP contributions between features. Additionally, it tackles a critical gap in feature selection literature by presenting criteria for the automatic feature selection of the most important features across a grid of models or stacked ensembles, eliminating the need for arbitrary determination of the number of top features to be extracted. This utility is invaluable for researchers analyzing feature significance, particularly within severely imbalanced outcomes where conventional methods fall short. Moreover, it is also expected to report democratic feature importance across a grid of models, resulting in a more comprehensive and generalizable feature selection. The package further implements a novel method for visualizing SHAP values both at subject level and feature level as well as a plot for feature selection based on the weighted mean SHAP ratios.

Maintained by E. F. Haghish. Last updated 13 days ago.

class-imbalance class-imbalance-problem feature-extraction feature-importance feature-selection machine-learning machine-learning-algorithms shap shap-analysis shap-values shapely shapley-additive-explanations shapley-decomposition shapley-value shapley-values shapleyvalue weighted-shap weighted-shap-confidence-interval weighted-shapley weighted-shapley-ci

0.5 match 15 stars 5.25 score 17 scripts