Sensitivity versus accuracy in ensemble models of Artificial Neural Networks from Multi-objective Evolutionary Algorithms

Research areas:

• Methodology - Machine learning

Year:

2018

Type of Publication:

Article

Keywords:

Ensemble, Multi-objective Evolutionary Algorithm, Multiclass classification, Artificial Neural Networks, Minimum Sensitivity, Pareto Performance measures

Authors:

• Fernández, Juan Carlos
• Cruz-Ramírez, Manuel
• Hervás-Martínez, César

Journal:

Neural Computing and Applications

Volume:

30

Number:

1

Pages:

289-305

Month:

June

ISSN:

0941-0643

BibTex:

@article{32016, author = "Juan Carlos Fern{\'a}ndez and Manuel Cruz-Ram{\'i}rez and C{\'e}sar Herv{\'a}s-Mart{\'i}nez", abstract = "This paper proposes a framework to obtain ensembles of classifiers from a Multi-objective Evolutionary Algorithm (MOEA), improving the restrictions imposed by two non-cooperative performance measures for multiclass problems: (1) the Correct Classification Rate or Accuracy (CCR) and, (2) the Minimum Sensitivity (MS) of all classes, i.e., the lowest percentage of examples correctly predicted as belonging to each class with respect to the total number of examples in the corresponding class. The proposed framework is based on collecting Pareto fronts of Artificial Neural Networks models for multiclass problems by the Memetic Pareto Evolutionary NSGA2 (MPENSGA2) algorithm, and it builds a new Pareto front (ensemble) from stored fronts. The ensemble built significantly improves the closeness to the optimum solutions and the diversity of the Pareto front. For verifying it, the performance of the new front obtained has been measured with the habitual use of weighting methodologies, such as Majority Voting, Simple Averaging and Winner Takes All. In addition to CCR and MS measures, three trade-off measures have been used to obtain the goodness of a Pareto front as a whole: Hyperarea, Laumanns’s Hyperarea (LAUMANNS) and Zitzler’s Spread (M3). The proposed framework can be adapted for any MOEA that aims to improve the compaction and diversity of its Pareto front, and whose fitness functions impose severe restrictions for multiclass problems.", awards = "JCR(2018): 4.664 Position: 21/133 (Q1) Category: COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE", comments = "JCR(2018): 4.664 Position: 21/133 (Q1) Category: COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE", doi = "http://dx.doi.org/10.1007/s00521-016-2781-y", issn = "0941-0643", journal = "Neural Computing and Applications", keywords = "Ensemble, Multi-objective Evolutionary Algorithm, Multiclass classification, Artificial Neural Networks, Minimum Sensitivity, Pareto Performance measures", month = "June", note = "JCR(2018): 4.664 Position: 21/133 (Q1) Category: COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE", number = "1", pages = "289-305", title = "{S}ensitivity versus accuracy in ensemble models of {A}rtificial {N}eural {N}etworks from {M}ulti-objective {E}volutionary {A}lgorithms", url = "http://dx.doi.org/10.1007/s00521-016-2781-y", volume = "30", year = "2018", }

Note:

JCR(2018): 4.664 Position: 21/133 (Q1) Category: COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Abstract:

This paper proposes a framework to obtain ensembles of classifiers from a Multi-objective Evolutionary Algorithm (MOEA), improving the restrictions imposed by two non-cooperative performance measures for multiclass problems: (1) the Correct Classification Rate or Accuracy (CCR) and, (2) the Minimum Sensitivity (MS) of all classes, i.e., the lowest percentage of examples correctly predicted as belonging to each class with respect to the total number of examples in the corresponding class. The proposed framework is based on collecting Pareto fronts of Artificial Neural Networks models for multiclass problems by the Memetic Pareto Evolutionary NSGA2 (MPENSGA2) algorithm, and it builds a new Pareto front (ensemble) from stored fronts. The ensemble built significantly improves the closeness to the optimum solutions and the diversity of the Pareto front. For verifying it, the performance of the new front obtained has been measured with the habitual use of weighting methodologies, such as Majority Voting, Simple Averaging and Winner Takes All. In addition to CCR and MS measures, three trade-off measures have been used to obtain the goodness of a Pareto front as a whole: Hyperarea, Laumanns’s Hyperarea (LAUMANNS) and Zitzler’s Spread (M3). The proposed framework can be adapted for any MOEA that aims to improve the compaction and diversity of its Pareto front, and whose fitness functions impose severe restrictions for multiclass problems.

Comments:

JCR(2018): 4.664 Position: 21/133 (Q1) Category: COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE