Algorithm Selection for Machine Learning Classification: An Application of the MELCHIOR Multicriteria Method
Igor Pinheiro de Araújo Costa, Marcio Pereira Basílio, Sérgio Mitihiro do Nascimento Maêda, Marcus Vinícius Gonçalves Rodrigues, Miguel Ângelo Lellis Moreira, Carlos Francisco Simões Gomes, Marcos dos Santos
This paper aims to select an algorithm for the Machine Learning (ML) classification task. For the proposed analysis, the Multi-criteria Decision Aid (MCDA) Méthode d’ELimination et de CHoix Includent les relations d’ORdre (MELCHIOR) method was applied. The experiment considered the following criteria as relevant: Accuracy, sensitivity, and processing time of the algorithms. The data used refers to the intention of buying on the Internet and the purpose is to predict whether the customer will finalize a particular purchase. Among various MCDA techniques available, MELCHIOR was chosen to support the decision-making process because this method provides the evaluation of alternatives without the need to elicit the weights of the criteria. As a result, the Gradient Boosting Decision Tree algorithm has been selected as the most suitable for the ML classification task.
IOS Press, Inc.
6751 Tepper Drive
Clifton, VA 20124
Tel.: +1 703 830 6300
Fax: +1 703 830 2300 firstname.lastname@example.org
(Corporate matters and books only) IOS Press c/o Accucoms US, Inc.
For North America Sales and Customer Service
West Point Commons
Lansdale PA 19446
Tel.: +1 866 855 8967
Fax: +1 215 660 5042 email@example.com