Belz, Julian: Fighting the curse of dimensionality with local model networks. 2018
Inhalt
- Titelblatt
- Acknowledgments
- Contents
- Symbols and Abbreviations
- Kurzfassung
- Abstract
- 1 Introduction
- 2 Nonlinear System Identification
- 2.1 Static and Dynamic Models
- 2.2 Curse of Dimensionality and Bias/Variance Tradeoff
- 2.3 Local Model Networks
- 2.4 Input Selection
- 2.5 Design of Experiments
- 2.6 Metamodeling
- 2.7 Static Function Generator
- 3 Input Selection Using Local Model Networks
- 3.1 Test Processes
- 3.2 Mixed Wrapper-Embedded Input Selection Approach
- 3.3 Regularization-Based Input Selection Approach
- 3.4 Embedded Approach
- 3.5 Visualization: Partial Dependence Plots
- 4 Design of Experiments Studies
- 4.1 Order Of Experimentation
- 4.2 Advisability of Specific Experimental Designs
- 4.3 Goal-Oriented Active Learning with Local Model Networks
- 5 Applications
- 5.1 Miles Per Gallon Data Set
- 5.2 Air-Mass Flow Prediction
- 5.3 Fan Metamodeling
- 5.4 Heating, Ventilating, and Air Conditioning System
- 6 Conclusions and Outlook
- A Data Splitting
- References