Pfannschmidt, Lukas: Relevance learning for redundant features. 2021
Inhalt
- Contents
- Introduction
- Foundations on Feature Selection
- Feature Relevance
- Feature Selection
- Feature Selection Methods for Possibly Redundant Features
- Summary
- Applications of Feature Relevance Bounds
- Ordinal Regression and the Relevance of Privileged Information
- Non-Linear Feature Selection and Classification
- Background
- Methods
- Loss-based Feature Set Decomposition
- Robust Loss Comparison
- Applications of Random Forest Importance Values
- Results
- Implementation
- Benchmark Models
- Stability of Feature Importance Values
- Parameterization for Feature Selection
- Linear Feature Selection Accuracy
- Non-Linear Feature Selection Accuracy
- Relevance Classification
- Conclusion
- Conclusion
- Appendices
- Appendix
- Relevance Bounds for Ordinal Regression
- Feature Relevance Bounds for Ordinal Regression with Implicit Order
- Proof of Generalization Bounds
- Proof of Theorem 1
- Equivalence of `3́9`42`"̇613A``45`47`"603AminRel() and `3́9`42`"̇613A``45`47`"603AminRel*()
- Equivalence of `3́9`42`"̇613A``45`47`"603AmaxRel() and the optimum of `3́9`42`"̇613A``45`47`"603AmaxRel*pos() and `3́9`42`"̇613A``45`47`"603AmaxRel*neg()
- Scaling of Ordinal Regression Feature Selection with Privileged Information
- Features of the COMPAS dataset
- Glossary
- Acronyms
- Bibliography
- List of Figures
- List of Tables