Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques
In this paper, we propose and compare new methodologies for ranking the importance of variables in productive processes via an adaptation of OneClass Support Vector Machines. In particular, we adapt two methodologies inspired by the machine learning literature: one involving the random shuffling of...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/11/11/2590 |
| _version_ | 1827739438409580544 |
|---|---|
| author | Raul Moragues Juan Aparicio Miriam Esteve |
| author_facet | Raul Moragues Juan Aparicio Miriam Esteve |
| author_sort | Raul Moragues |
| collection | DOAJ |
| description | In this paper, we propose and compare new methodologies for ranking the importance of variables in productive processes via an adaptation of OneClass Support Vector Machines. In particular, we adapt two methodologies inspired by the machine learning literature: one involving the random shuffling of values of a variable and another one using the objective value of the dual formulation of the model. Additionally, we motivate the use of these type of algorithms in the production context and compare their performance via a computational experiment. We observe that the methodology based on shuffling the values of a variable outperforms the methodology based on the dual formulation. We observe that the shuffling-based methodology correctly ranks the variables in 94% of the scenarios with one relevant input and one irrelevant input. Moreover, it correctly ranks each variable in at least 65% of replications of a scenario with three relevant inputs and one irrelevant input. |
| first_indexed | 2024-03-11T03:01:47Z |
| format | Article |
| id | doaj.art-8c172dfe59064707979c03e6e6b9b4de |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-11T03:01:47Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-8c172dfe59064707979c03e6e6b9b4de2023-11-18T08:14:07ZengMDPI AGMathematics2227-73902023-06-011111259010.3390/math11112590Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning TechniquesRaul Moragues0Juan Aparicio1Miriam Esteve2Center of Operations Research (CIO), Miguel Hernandez University of Elche (UMH), 03202 Elche, SpainCenter of Operations Research (CIO), Miguel Hernandez University of Elche (UMH), 03202 Elche, SpainCenter of Operations Research (CIO), Miguel Hernandez University of Elche (UMH), 03202 Elche, SpainIn this paper, we propose and compare new methodologies for ranking the importance of variables in productive processes via an adaptation of OneClass Support Vector Machines. In particular, we adapt two methodologies inspired by the machine learning literature: one involving the random shuffling of values of a variable and another one using the objective value of the dual formulation of the model. Additionally, we motivate the use of these type of algorithms in the production context and compare their performance via a computational experiment. We observe that the methodology based on shuffling the values of a variable outperforms the methodology based on the dual formulation. We observe that the shuffling-based methodology correctly ranks the variables in 94% of the scenarios with one relevant input and one irrelevant input. Moreover, it correctly ranks each variable in at least 65% of replications of a scenario with three relevant inputs and one irrelevant input.https://www.mdpi.com/2227-7390/11/11/2590data envelopment analysisfeature rankingmodel specificationunsupervised machine learningtechnical efficiencyoverfitting |
| spellingShingle | Raul Moragues Juan Aparicio Miriam Esteve Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques Mathematics data envelopment analysis feature ranking model specification unsupervised machine learning technical efficiency overfitting |
| title | Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques |
| title_full | Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques |
| title_fullStr | Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques |
| title_full_unstemmed | Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques |
| title_short | Ranking the Importance of Variables in a Nonparametric Frontier Analysis Using Unsupervised Machine Learning Techniques |
| title_sort | ranking the importance of variables in a nonparametric frontier analysis using unsupervised machine learning techniques |
| topic | data envelopment analysis feature ranking model specification unsupervised machine learning technical efficiency overfitting |
| url | https://www.mdpi.com/2227-7390/11/11/2590 |
| work_keys_str_mv | AT raulmoragues rankingtheimportanceofvariablesinanonparametricfrontieranalysisusingunsupervisedmachinelearningtechniques AT juanaparicio rankingtheimportanceofvariablesinanonparametricfrontieranalysisusingunsupervisedmachinelearningtechniques AT miriamesteve rankingtheimportanceofvariablesinanonparametricfrontieranalysisusingunsupervisedmachinelearningtechniques |