Abstract
First, we explore the properties of families of odd-point odd-ary parametric approximating subdivision schemes. Then we fine-tune the parameters involved in the family of schemes to maximize the smoothness of the limit curve and error bounds for the distance between the limit curve and the kth level control polygon. After that, we present the subdivision-regularization framework for preventing over fitting of data by model. Demonstration shows that the proposed unified frame work can work well for both noise removal and overfitting prevention in subdivision as well as regularization.
Recommended Citation
Mustafa, Ghulam; Ghaffar, Abdul; and Aslam, Muhammad
(2013).
A Subdivision-Regularization Framework for Preventing Over Fitting of Data by a Model,
Applications and Applied Mathematics: An International Journal (AAM), Vol. 8,
Iss.
1, Article 11.
Available at:
https://digitalcommons.pvamu.edu/aam/vol8/iss1/11