A family of smooth piecewise-linear models with probabilistic interpretations

The smooth piecewise-linear models cover a wide range of applications nowadays. Basically, there are two classes of them: models are transitional or hyperbolic according to their behaviour at the phase-transition zones. This study explored three different approaches to build smooth piecewise-linear models, and we analysed their inter-relationships by a unifying modelling framework. We conceived the smoothed phase-transition zones as domains where a mixture process takes place, which ensured probabilistic interpretations for both hyperbolic and transitional models in the light of random thresholds. Many popular models found in the literature are special cases of our methodology. Furthermore, this study introduces novel regression models as alternatives, such as the Epanechnikov, Normal and Skewed-Normal Bent-Cables.

[1]  D. Griffiths,et al.  Hyperbolic regression - a model based on two-phase piecewise linear regression with a smooth transition between regimes , 1973 .

[2]  V. Muggeo.  Estimating regression models with unknown break‐points , 2003, Statistics in medicine.

[3]  J. Hanna,et al.  Letters to the Editor , 1999, Journal of paediatrics and child health.

[4]  M. Lesperance,et al.  PIECEWISE REGRESSION: A TOOL FOR IDENTIFYING ECOLOGICAL THRESHOLDS , 2003 .

[5]  Exponentiated uniform distribution: An interesting alternative to truncated models , 2019 .

[6]  Andrej Pázman,et al.  Nonlinear Regression , 2019, Handbook of Regression Analysis With Applications in R.

[7]  Israel Zang,et al.  A smoothing-out technique for min—max optimization , 1980, Math. Program..

[8]  Asher Tishler,et al.  A Maximum Likelihood Method for Piecewise Regression Models with a Continuous Dependent Variable , 1981 .

[9]  Shahedul A. Khan,et al.  Generalized bent-cable methodology for changepoint data: a Bayesian approach , 2018 .

[10]  Peter Berck,et al.  Reconciling the von Liebig and Differentiable Crop Production Functions , 1990 .

[11]  Relationship between zooplankton richness and area in Brazilian lakes: comparing natural and artificial lakes and trends , 2018, Acta Limnologica Brasiliensia.

[12]  Iuri Emmanuel de Paula Ferreira,et al.  Reconciling the Mitscherlich's law of diminishing returns with Liebig's law of the minimum. Some results on crop modeling. , 2017, Mathematical biosciences.

[13]  Grace S. Chiu,et al.  Bent-Cable Regression Theory and Applications , 2006 .

[14]  A. Blackmer,et al.  Comparison of Models for Describing; Corn Yield Response to Nitrogen Fertilizer , 1990 .

[15]  J.-N. Lin,et al.  Canonical representation: from piecewise-linear function to piecewise-smooth functions , 1993 .

[16]  David W. Bacon,et al.  Using An Hyperbola as a Transition Model to Fit Two-Regime Straight-Line Data , 1974 .

[17]  David W. Bacon,et al.  Estimating the transition between two intersecting straight lines , 1971 .

[18]  M. Muggeo,et al.  segmented: An R package to Fit Regression Models with Broken-Line Relationships , 2008 .

[19]  H. Jacqmin-Gadda,et al.  A curvilinear bivariate random changepoint model to assess temporal order of markers , 2020, Statistical methods in medical research.

[20]  A. Tishler,et al.  A New Maximum Likelihood Algorithm for Piecewise Regression , 1981 .

[21]  Lyle Prunty,et al.  Curve fitting with smooth functions that are piecewise-linear in the limit , 1983 .

[22]  D. Bertsekas.  Nondifferentiable optimization via approximation , 1975 .

[23]  C. Pantaleon,et al.  Smoothing the canonical piecewise-linear model: an efficient and derivable large-signal model for MESFET/HEMT transistors , 2001 .

[24]  Irina Roslyakova,et al.  Modeling thermodynamical properties by segmented non-linear regression , 2017 .

[25]  H. H. Cerecedo-Núñez,et al.  Transforming the canonical piecewise-linear model into a smooth-piecewise representation , 2016, SpringerPlus.