Use of generalized additive mixed models in the comparison of mortar types in different configurations

• Autor(a) principal: Silva, Breno Gabriel da
• Data de Publicação: 2022
• Outros Autores: Oliveira, Willian Luís de, Guedes, Terezinha Aparecida, Janeiro, Vanderly, Canova, José Aparecido
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Research, Society and Development
• DOI: 10.33448/rsd-v11i6.29160
• Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/29160
• topic: Modelos aditivos generalizados mistos; Dados longitudinais; Regressão não paramétrica; P-splines.; Modelos aditivos generalizados mixtos; Datos longitudinales; Regresión no paramétrica; P-splines.; Generalized additive mixed models; Longitudinal data; Nonparametric regression; P-splines.

Generalized additive mixed models (GAMM) are useful when the nature of the data is longitudinal, which include the nonlinearity of the individual trajectories of the subjects, associating these with the assumption of the existence of random effects for each individual. In these models it is possible to rewrite the predictor as a sum of smooth nonparametric functions and then use the smooth technique P-spline. Thus, using the generalized additive mixed models with P-splines, this article aims to verify the impact of the types of mortars and the concentration levels of an additive content on the evolution of mortar weight contained in specimens over time, verifying possible differences between combinations of types and concentration. In relation to the results, it was observed through the analyzes that there were significant differences in the estimated curves of the weight evolution in both types of mortar, concluding that the matured mortar has a water absorption speed by capillarity higher than dry mortar. All the necessary assumptions for the validity of the model have been satisfied.