Topic
I-spline
About: I-spline is a research topic. Over the lifetime, 5 publications have been published within this topic receiving 72 citations.
Papers
1 Jan 2012
TL;DR: Numerical studies based on three data sets showed the empirical evidences of I-spline Smoothing in improving calibration in improving calibrating predictive models without deterioration of discrimination.
Abstract: We proposed the I-spline Smoothing approach for calibrating predictive models by solving a nonlinear monotone regression problem. We took advantage of I-spline properties to obtain globally optimal solutions while keeping the computational cost low. Numerical studies based on three data sets showed the empirical evidences of I-spline Smoothing in improving calibration (i.e.,1.6x, 1.4x, and 1.4x on the three datasets compared to the average of competitors-Binning, Platt Scaling, Isotonic Regression, Monotone Spline Smoothing, Smooth Isotonic Regression) without deterioration of discrimination.
5 citations
TL;DR: R Ramsay's algorithm for estimating monotonic transformations in regression is modified and extended and can capture some characteristics that are pertinent to the time series and is much easier to implement.
Abstract: Summary. We modify Ramsay's algorithm for estimating monotonic transformations in regression and extend it to autoregression, where strict monotonicity is an essential requirement. Compared with other methods, our method can capture some characteristics that are pertinent to the time series and is much easier to implement. An order selection method is introduced and developed. Some real data sets are analysed.
5 citations
TL;DR: Extensive numerical studies show that the proposed penalized regression spline estimator captures spatially inhomogeneous behaviors of data, such as sudden jumps.
Abstract: We propose a penalized regression spline estimator for monotone regression. To construct the estimator, we adopt the I-splines with the total variation penalty. The I-splines lend themselves to the...
3 citations
TL;DR: The I-spline approximation of the link function via maximizing the covariance function with a penalty function is investigated in the present work and the consistency of the criterion is constructed.
Abstract: The single-index model with monotonic link function is investigated. Firstly, it is showed that the link function h(·) can be viewed by a graphic method. That is, the plot with the fitted response ŷ on the horizontal axis and the observed y on the vertical axis can be used to visualize the link function. It is pointed out that this graphic approach is also applicable even when the link function is not monotonic. Note that many existing nonparametric smoothers can also be used to assess h(·). Therefore, the I-spline approximation of the link function via maximizing the covariance function with a penalty function is investigated in the present work. The consistency of the criterion is constructed. A small simulation is carried out to evidence the efficiency of the approach proposed in the paper.
Book•
1 Jan 2003
TL;DR: In this article, non-parametric spline models are used to detect filamentary entities and to reconstruct and match filamentary structures. But they are not suitable for modeling in action.
Abstract: Acknowledgments*List of Figures*Notation and Symbols*Introduction * I Spline Models*Non-parametric spline models * Parametric spline models * Auto-Associative Models * II Markov Models*Fundamental Aspects * Bayesian estimation * Simulation and optimization * Parameter Estimation *III Modeling in Action* Model-building * Degradation in Imaging * Detection of filamentary Entities * Reconstruction and Projections * Matching * References
Related Topics (5)
Performance Metrics
| Year | Papers |
|---|---|
| 2021 | 1 |
| 2012 | 1 |
| 2008 | 1 |
| 2003 | 1 |
| 2000 | 1 |