University of Twente Student Theses
Role of Classifier in Feature Importance : An Empirical Evaluation
Meng, Huanbo (2025) Role of Classifier in Feature Importance : An Empirical Evaluation.
|
PDF
817kB |
| Abstract: | Feature importance methods, such as SHAP (SHapley Additive exPlanations) and permutation feature importance, are widely utilized to assess the influence of features on model outcomes. In theory, these methods should provide consistent importance rankings regardless of the classifier employed. However, the internal mechanisms of classifiers can significantly influence the resulting importance scores in practice. This research aims to investigate the impact of four feature importance measurements on five different classifiers across four datasets. Comparative analyses of feature importance ranking results will be conducted to identify inconsistencies and patterns linked to classifier choice. We can infer from the research that feature importance techniques SHAP and PFI are more closely related to linear classifiers. Feature importance ranking results are mostly influenced by the generalizability of features and the linearity of the classifier. Additionally, one can alter the ranking results to achieve more desirable results by adjusting the hyperparameter. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/105157 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page