Busca avançada
Ano de início
(Referência obtida automaticamente do Web of Science, por meio da informação sobre o financiamento pela FAPESP e o número do processo correspondente, incluída na publicação pelos autores.)

A meta-learning recommender system for hyperparameter tuning: Predicting when tuning improves SVM classifiers

Texto completo
Mantovani, Rafael G. [1, 2] ; Rossi, Andre L. D. [3] ; Alcobaca, Edesio [1] ; Vanschoren, Joaquin [4] ; de Carvalho, Andre C. P. L. F. [1]
Número total de Autores: 5
Afiliação do(s) autor(es):
[1] Univ Sao Paulo, Inst Math & Comp Sci, Sao Carlos, SP - Brazil
[2] Fed Technol Univ, Campus Apucarana, R Marcilio Dias 635, BR-86812460 Apucarana, PR - Brazil
[3] Univ Estadual Paulista, Campus Itapeva, Sao Paulo - Brazil
[4] Eindhoven Univ Technol, Eindhoven - Netherlands
Número total de Afiliações: 4
Tipo de documento: Artigo Científico
Fonte: INFORMATION SCIENCES; v. 501, p. 193-221, OCT 2019.
Citações Web of Science: 0

For many machine learning algorithms, predictive performance is critically affected by the hyperparameter values used to train them. However, tuning these hyperparameters can come at a high computational cost, especially on larger datasets, while the tuned settings do not always significantly outperform the default values. This paper proposes a recommender system based on meta-learning to identify exactly when it is better to use default values and when to tune hyperparameters for each new dataset. Besides, an in-depth analysis is performed to understand what they take into account for their decisions, providing useful insights. An extensive analysis of different categories of meta-features, meta-learners, and setups across 156 datasets is performed. Results show that it is possible to accurately predict when tuning will significantly improve the performance of the induced models. The proposed system reduces the time spent on optimization processes, without reducing the predictive performance of the induced models (when compared with the ones obtained using tuned hyperparameters). We also explain the decision-making process of the meta-learners in terms of linear separability-based hypotheses. Although this analysis is focused on the tuning of Support Vector Machines, it can also be applied to other algorithms, as shown in experiments performed with decision trees. (C) 2019 Published by Elsevier Inc. (AU)

Processo FAPESP: 12/23114-9 - Uso de meta-aprendizado para ajuste de parâmetros em problemas de classificação
Beneficiário:Rafael Gomes Mantovani
Linha de fomento: Bolsas no Brasil - Doutorado
Processo FAPESP: 15/03986-0 - Uso de Meta-aprendizado para melhoria de algoritmos de deep learning em problemas de classificação
Beneficiário:Rafael Gomes Mantovani
Linha de fomento: Bolsas no Exterior - Estágio de Pesquisa - Doutorado
Processo FAPESP: 18/14819-5 - Aprendizado de máquina automático: aprendendo a aprender
Beneficiário:Edesio Pinto de Souza Alcobaça Neto
Linha de fomento: Bolsas no Brasil - Doutorado Direto