MetaSel 2014
Meta-learning and Algorithm Selection

Table of Contents

Preface

Invited Talk Abstracts

• Towards an Algorithm Selection Standard: Data Format and Tools1-1
  Lars Kotthoff
  
• Bayesian Optimization for More Automatic Machine Learning2-2
  Frank Hutter
  

Technical Papers

• Using Meta-Learning to Initialize Bayesian Optimization of Hyperparameters3-10
  Matthias Feurer, Tobias Springenberg, Frank Hutter
  
• Similarity Measures of Algorithm Performance for Cost-Sensitive Scenarios11-17
  Carlos Eduardo Castor de Melo, Ricardo Prudêncio
  
• Using Metalearning to Predict When Parameter Optimization Is Likely to Improve Classification Accuracy18-23
  Parker Ridd, Christophe Giraud-Carrier
  
• Surrogate Benchmarks for Hyperparameter Optimization24-31
  Katharina Eggensperger, Frank Hutter, Holger Hoos, Kevin Leyton-Brown
  
• A Framework To Decompose And Develop Metafeatures32-36
  Fabio Pinto, Carlos Soares, Joao Mendes-Moreira
  
• Towards Meta-learning over Data Streams37-38
  Jan van Rijn, Geoffrey Holmes, Bernhard Pfahringer, Joaquin Vanschoren
  
• Recommending Learning Algorithms and Their Associated Hyperparameters39-40
  Michael Smith, Logan Mitchell, Christophe Giraud-Carrier, Tony Martinez
  
• An Easy to Use Repository for Comparing and Improving Machine Learning Algorithm Usage41-48
  Michael Smith, Andrew White, Christophe Giraud-Carrier, Tony Martinez
  
• Measures for Combining Accuracy and Time for Meta-learning49-50
  Salisu Abdulrahman, Pavel Brazdil
  

Abstracts

• Determining a Proper Initial Configuration of Red-Black Planning by Machine Learning51-52
  Otakar Trunda, Roman Bartak
  
• Hybrid Multi-Agent System for Metalearning in Data Mining53-54
  Klara Peskova, Jakub Smid, Martin Pilat, Ondrej Kazik, Roman Neruda
  
• Model Selection in Data Analysis Competitions55-60
  David Kofoed Wind, Ole Winther