Universiti Teknologi Malaysia Institutional Repository

Enhanced nadaraya-watson kernel surface approximation for extremely small samples

Shapiai @ Abd. R., Mohd. Ibrahim and Ibrahim, Zuwairie and Khalid, Marzuki and Lee, Jau Wen and Pavlovich, Vladimir (2011) Enhanced nadaraya-watson kernel surface approximation for extremely small samples. In: The 5th Asia Asia Modelling Symposium 2011 (Ams 2011).

Full text not available from this repository.

Official URL: http://dx.doi.org/10.1109/AMS.2011.13


The function approximation problem is to find the appropriate relationship between a dependent and independent variable(s). Function approximation algorithms generally require sufficient samples to approximate a function. Insufficient samples may cause any approximation algorithm to result in unsatisfactory predictions. To solve this problem, a function approximation algorithm called Weighted Kernel Regression (WKR), which is based on Nadaraya-Watson kernel regression, is proposed. In the proposed framework, the original Nadaraya-Watson kernel regression algorithm is enhanced by expressing the observed samples in a square kernel matrix. The WKR is trained to estimate the weight for the testing phase. The weight is estimated iteratively and is governed by the error function to find a good approximation model. Two experiments are conducted to show the capability of the WKR. The results show that the proposed WKR model is effective in cases where the target surface function is non-linear and the given training sample is small. The performance of the WKR is also compared with other existing function approximation algorithms, such as artificial neural networks (ANN).

Item Type:Conference or Workshop Item (Paper)
Uncontrolled Keywords:kernel surface
Divisions:Electrical Engineering
ID Code:45824
Deposited By: Haliza Zainal
Deposited On:10 Jun 2015 03:01
Last Modified:29 Aug 2017 03:58

Repository Staff Only: item control page