Premium
Smoothing using fractional polynomials: an alternative to polynomials and splines in applied research
Author(s) -
Regier Michael David,
Parker R. David
Publication year - 2015
Publication title -
wiley interdisciplinary reviews: computational statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.693
H-Index - 38
eISSN - 1939-0068
pISSN - 1939-5108
DOI - 10.1002/wics.1355
Subject(s) - polynomial , curvilinear coordinates , interpretation (philosophy) , mathematics , smoothing , polynomial regression , computer science , algebra over a field , regression analysis , calculus (dental) , statistics , pure mathematics , mathematical analysis , geometry , programming language , medicine , dentistry
The fractional polynomial regression model is an emerging tool in applied research. Overcoming inherent problems associated with a polynomial expansion and splines, fractional polynomial models provide an alternate approach for modeling nonlinear relationships. In this article, we introduce the univariable and multivariable fractional polynomial model and highlight important aspects of their construction. Because of the curvilinear nature of fractional polynomial models, functional tables and functional plots are emphasized for model interpretation. We present two examples to illustrate fractional polynomial models for their selection and interpretation in applied research. WIREs Comput Stat 2015, 7:275–283. doi: 10.1002/wics.1355 This article is categorized under: Statistical and Graphical Methods of Data Analysis > Density Estimation Statistical Learning and Exploratory Methods of the Data Sciences > Modeling Methods Statistical and Graphical Methods of Data Analysis > Transformations