Javascript must be enabled to continue!
Greedy knot selection algorithm for restricted cubic spline regression
View through CrossRef
Non-linear regression modeling is common in epidemiology for prediction purposes or estimating relationships between predictor and response variables. Restricted cubic spline (RCS) regression is one such method, for example, highly relevant to Cox proportional hazard regression model analysis. RCS regression uses third-order polynomials joined at knot points to model non-linear relationships. The standard approach is to place knots by a regular sequence of quantiles between the outer boundaries. A regression curve can easily be fitted to the sample using a relatively high number of knots. The problem is then overfitting, where a regression model has a good fit to the given sample but does not generalize well to other samples. A low knot count is thus preferred. However, the standard knot selection process can lead to underperformance in the sparser regions of the predictor variable, especially when using a low number of knots. It can also lead to overfitting in the denser regions. We present a simple greedy search algorithm using a backward method for knot selection that shows reduced prediction error and Bayesian information criterion scores compared to the standard knot selection process in simulation experiments. We have implemented the algorithm as part of an open-source R-package,
knutar
.
Frontiers Media SA
Title: Greedy knot selection algorithm for restricted cubic spline regression
Description:
Non-linear regression modeling is common in epidemiology for prediction purposes or estimating relationships between predictor and response variables.
Restricted cubic spline (RCS) regression is one such method, for example, highly relevant to Cox proportional hazard regression model analysis.
RCS regression uses third-order polynomials joined at knot points to model non-linear relationships.
The standard approach is to place knots by a regular sequence of quantiles between the outer boundaries.
A regression curve can easily be fitted to the sample using a relatively high number of knots.
The problem is then overfitting, where a regression model has a good fit to the given sample but does not generalize well to other samples.
A low knot count is thus preferred.
However, the standard knot selection process can lead to underperformance in the sparser regions of the predictor variable, especially when using a low number of knots.
It can also lead to overfitting in the denser regions.
We present a simple greedy search algorithm using a backward method for knot selection that shows reduced prediction error and Bayesian information criterion scores compared to the standard knot selection process in simulation experiments.
We have implemented the algorithm as part of an open-source R-package,
knutar
.
Related Results
Greedy Knot Selection Algorithm for Restricted Cubic Spline Regression
Greedy Knot Selection Algorithm for Restricted Cubic Spline Regression
Abstract
Non-linear regression modeling is common in many fields for prediction purposes or estimating
relationships between predictor and response variables. For e...
Greedy Knot Selection Algorithm for Restricted Cubic Spline Regression
Greedy Knot Selection Algorithm for Restricted Cubic Spline Regression
Abstract
Non-linear regression modeling is common in many fields for prediction purposes or estimating relationships between predictor and respon...
Greedy Knot Selection Algorithm for Restricted Cubic Spline Regression
Greedy Knot Selection Algorithm for Restricted Cubic Spline Regression
Abstract
Non-linear regression modeling is common in many fields for prediction purposes or estimating
relationships between predictor and response variables. For e...
Towards an Improved Strategy for Solving Multi-Armed Bandit Problem
Towards an Improved Strategy for Solving Multi-Armed Bandit Problem
Multi-Armed Bandit (MAB) problem is one of the classical reinforcements learning problems that describe the friction between the agent’s exploration and exploitation. This study ex...
Extensive Experimental Study on the Stability of Rotor System With Spline Coupling
Extensive Experimental Study on the Stability of Rotor System With Spline Coupling
Spline couplings which have simple structure, high reliability and can compensate torque transmission error are widely used in rotating machineries, such as aeroengine and gasturbi...
Vindumhede-fletningerne og kjærlighetsknop
Vindumhede-fletningerne og kjærlighetsknop
How and Why the Knot in the Vindumhede Plait was tied - and two Related Norwegian KnotsIn ancient times and in the Middle Ages knots were believed to have magic or mystic power. S...
Vessel Trajectory Similarity Based on Cubic B-spline Interpolation + Time-constrained Hausdorff Algorithm
Vessel Trajectory Similarity Based on Cubic B-spline Interpolation + Time-constrained Hausdorff Algorithm
Abstract
The spatio-temporal trajectory data sampling period is large, and the general trajectory similarity is not suitable. This paper proposes a new algorithm, cu...

