CPMP-2015-runtime-classification
OpenML dataset with id 41919
Author name not available (Why is that?)
Full work available at URL: https://api.openml.org/data/v1/download/21378712/CPMP-2015-runtime-classification.arff
Upload date: 19 June 2019
Dataset Characteristics
Number of classes: 4
Number of features: 23 (numeric: 22, symbolic: 1 and in total binary: 0 )
Number of instances: 527
Number of instances with missing values: 0
Number of missing values: 0
source: An Algorithm Selection Benchmark for the Container Pre-Marshalling Problem (CPMP) authors: K. Tierney and Y. Malitsky (features) / K. Tierney and D. Pacino and S. Voss (algorithms) translator in coseal format: K. Tierney
This is an extension of the 2013 premarshalling dataset that includes more features and a set of test instances.
There are three sets of features:
feature_values.arff contains the full set of features from iteration 2 of our latent feature analysis (LFA) process (see paper) feature_values_itr1.arff contains only the features after iteration 1 of LFA feature_values_orig.arff containers the features used in PREMARHSALLING-ASTAR-2013
We also provide test data with an identical naming scheme (see _test).
The features for the pre-marshalling problem are all extremely easy and fast to compute, thus the feature_costs.arff file has been omitted, as it would be time 0 for every feature (regardless of using original, iteration 1 or iteration 2 features).
The feature computation code is available at https://bitbucket.org/eusorpb/cpmp-as
Note: previously the scenario was called PREMARSHALLING-ASTAR-2015. To save same space, we renamed the scenario.
This page was built for dataset: CPMP-2015-runtime-classification
