An Evaluation of Landmarking Variants

By Johannes Fürnkranz, Johann Petrak. 2001

98 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Automated Machine Learning: State-of-The-Art and Open Challenges (2019, methodology): “One way to measure the similarity between tnew and tj is using the relative landmarks that measures the performance difference between two model configurations on the same task [42].”
Meta-Learning: A Survey (2018, neutral): “A first measure for task similarity considers the relative (pairwise) performance differences, also called relative landmarks, RLa,b,j = Pa,j − Pb,j between two configurations θa and θb on a particular task tj (Fürnkranz and Petrak, 2001).”
On the predictive power of meta-features in OpenML (2017, background): “Landmarking measures have been evaluated and have shown to perform well in many works (Fürnkranz and Petrak, 2001; Bensusan and Kalousis, 2001; Pfahringer et al., 2000; Reif et al., 2014).”
Ranking Learning Algorithms: Using IBL and Meta-Learning on Accuracy and Time Results (2003, methodology): “Landmarks are quick estimates of algorithm performance on a given dataset obtained using simple versions of the algorithms (Bensusan & Giraud-Carrier, 2000; Pfahringer, Bensusan, & Giraud-Carrier, 2000) or by sampling from the dataset (F¨urnkranz & Petrak, 2001; Soares, Petrak, & Brazdil, 2001b).”
Using Meta-Learning to Support Data Mining (2004, methodology): “This information is subsequently used to select an appropriate learning algorithm ([24],[45]).”
Cross-disciplinary perspectives on meta-learning for algorithm selection (2009, methodology): “…as an alternative to generating computationally expensive features was also developed further during METAL to create models predicting the performance of algorithms relative to the performance of simpler and faster learners [Pfahringer et al. 2000; Furnkranz and Petrak 2001; Soares et al. 2001].”
Automatic classifier selection for non-experts (2014, methodology): “Theoretic ModelBased Landmarking Gama and Brazdil [17] • • • Köpf et al. [24] • • • Bensusan and Kalousis [7] • • • • a1 Linder et al. [25] • • • Brazdil et al. [8] • • • m Ali et al. [2] • • • • Fürnkranz and Petrak [16] • Pfahringer et al. [30] • • Bensuan et al. [4] • • Peng et al. [29] • • • •”
Meta-QSAR: a large-scale application of meta-learning to drug design and discovery (2017, methodology): “running all candidate algorithms on several small samples of the new data (Fürnkranz and Petrak, 2001).”
Fast Algorithm Selection Using Learning Curves (2015, methodology): “The authors of [4] show that this information is not useful as landmarker.”
A review of automatic selection methods for machine learning algorithms and hyper-parameter values (2016, methodology): “Sampling-based landmark (Petrak 2000; Fürnkranz and Petrak 2001; Soares et al. 2001) is another method for automatically selecting machine learning algorithms.”