Multivariate Adaptive Regression Splines (MARS) for non-linear time series dependence: An application in biostatistics
Over several years, time series analysis using ARIMA modelling has been proposed as a suitable method to investigate the nonlinear relationship between antimicrobial use and resistance. Nonetheless, these studies did not mention the thresholds of antibiotic use to manage the precise prescription. Th...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2021 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/348071 |
| Acceso en línea: | https://hdl.handle.net/2117/348071 |
| Access Level: | acceso abierto |
| Palabra clave: | Statistical mechanics MARS TF Antimicrobial use and resistance rate Mecànica estadística Classificació AMS::82 Statistical mechanics, structure of matter::82C Time-dependent statistical mechanics (dynamic and nonequilibrium) Àrees temàtiques de la UPC::Matemàtiques i estadística::Estadística matemàtica |
| Sumario: | Over several years, time series analysis using ARIMA modelling has been proposed as a suitable method to investigate the nonlinear relationship between antimicrobial use and resistance. Nonetheless, these studies did not mention the thresholds of antibiotic use to manage the precise prescription. This paper explores using multivariate adaptive regression splines for nonlinear time series to develop models applying monthly antimicrobial resistance and antimicrobial use data. And according to the study of the MARS methodology with the application to antibiotic use and resistance, this type of model is proved to be useful to identify relationships between explanatory and outcome time series. The approach is very flexible and allows for the inclusion of smooth functions, thresholds and lags in the input time series that reflect non-linear dependencies. |
|---|