Skip to main content
A Kernel Path Algorithm For General Parametric Quadratic Programming Problem
Pattern Recognition
  • Bin Gu, Nanjing University of Information Science & Technology Nanjing, PR China & Mohamed bin Zayed University of Artificial Intelligence
  • Ziran Xiong, Nanjing University of Information Science & Technology
  • Shuyang Yu, Northeastern University, Shengyang, PR China
  • Guansheng Zheng, Nanjing University of Information Science & Technology, Nanjing, PR China
Document Type

It is well known that the performance of a kernel method highly depends on the choice of kernel parameter. A kernel path provides a compact representation of all optimal solutions, which can be used to choose the optimal value of kernel parameter along with cross validation (CV) method. However, none of these existing kernel path algorithms provides a unified implementation to various learning problems. To fill this gap, in this paper, we first study a general parametric quadratic programming (PQP) problem that can be instantiated to an extensive number of learning problems. Then we provide a generalized kernel path (GKP) for the general PQP problem. Furthermore, we analyze the iteration complexity and computational complexity of GKP. Extensive experimental results on various benchmark datasets not only confirm the identity of GKP with several existing kernel path algorithms, but also show that our GKP is superior to the existing kernel path algorithms in terms of generalization and robustness.

Publication Date
  • Cross validation,
  • Kernel path,
  • Parametric quadratic programming,
  • QR decomposition

IR deposit conditions:

  • OA (accepted version) - pathway b
  • 24 months embargo
  • Must link to publisher version with DOI
Citation Information
B. Gu, Z. Xiong, S. Yu, and G. Zheng, “A kernel path algorithm for general parametric quadratic programming problem,” Pattern Recognition, vol. 116, p. 2-10, Aug. 2021, doi: 10.1016/J.PATCOG.2021.107941.