Predicting Algorithm Accuracy with a Small Set of Effective Meta-Features

We revisit 26 meta-features typically used in the context of meta-learning for model selection. Using visual analysis and computational complexity considerations, we find 4 meta-features whose values are directly relevant to certain ranges of predictive accuracy for 7 learning algorithms on 135 UCI datasets. Discretization of these 4 meta-features based on thresholds derived from our analysis significantly boosts the accuracy of the meta-level classification task.

• Type: Conference
• Year: 2008
• Where: ICMLA
• Authors: Jun Won Lee, Christophe G. Giraud-Carrier
• Keywords: Computational Complexity Considerations, ICMLA 2008, Machine Learning, Meta-level Classification Task, Predictive Accuracy