Function and surface approximation based on enhanced Kernel Regression for small sample sets

Mohd Ibrahim Shapiai, Zuwairie Ibrahim, Marzuki Khalid, Lee Wen Jau, Vladimir Pavlovic, Junzo Watada

    Research output: Contribution to journalArticlepeer-review

    8 Citations (Scopus)


    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 function 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 (NWKR), is proposed. In the proposed framework, the original NWKR 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 governed by the error function to find a good approximation model. Four 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 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).

    Original languageEnglish
    Pages (from-to)5947-5960
    Number of pages14
    JournalInternational Journal of Innovative Computing, Information and Control
    Issue number10
    Publication statusPublished - 2011 Oct


    • Artificial neural network
    • Non-linear function
    • Small samples
    • Weighted kernel regression

    ASJC Scopus subject areas

    • Computational Theory and Mathematics
    • Information Systems
    • Software
    • Theoretical Computer Science


    Dive into the research topics of 'Function and surface approximation based on enhanced Kernel Regression for small sample sets'. Together they form a unique fingerprint.

    Cite this