38 A review of: Effect of the Sampling of a Dataset in the Hyperparameter Optimization Phase over the Efficiency of a Machine Learning Algorithm

Cybersecurity is a discipline in which artificial intelligence techniques are gaining in importance in order to obtain actionable knowledge. The selection of a fitting configuration of hyperparameters is an important factor in model performance. Several hyperparameter optimization algoritms have bee...

Descripción completa

Detalles Bibliográficos
Autores: Castro-García, Noemí de, Muñoz Castañeda, Ángel Luis, Escudero García, David
Tipo de recurso: capítulo de libro
Fecha de publicación:2021
País:España
Institución:Universidad de Castilla-La Mancha
Repositorio:RUIdeRA. Repositorio Institucional de la UCLM
OAI Identifier:oai:ruidera.uclm.es:10578/28648
Acceso en línea:http://doi.org/10.18239/jornadas_2021.34.38
http://hdl.handle.net/10578/28648
Access Level:acceso abierto
Palabra clave:cybersegurity
Machine Learning
Hyperparameter optimization
Descripción
Sumario:Cybersecurity is a discipline in which artificial intelligence techniques are gaining in importance in order to obtain actionable knowledge. The selection of a fitting configuration of hyperparameters is an important factor in model performance. Several hyperparameter optimization algoritms have been developed, but their application imposes additional computational costs. Since one of the main factors in the resource consumption is the dataset size, we perform a study of the effect of using different partitions over five different cybersecurity datasets. Nonparametric inference has been used to measure the rate of change of the accuracy, time, and spatial (memory) complexity along the partition size. In addition, a level of gain is assigned to each partition allowing us to study patterns and determine the optimal partition size.