A Greedy Algorithm for Constraint Principal Curves

Date

2014

Authors

Yang, S.
Chen, D.
Zeng, X.
Pudney, P.

Editors

Advisors

Journal Title

Journal ISSN

Volume Title

Type:

Journal article

Citation

Journal of Computers, 2014; 9(5):1125-1130

Statement of Responsibility

Conference Name

Abstract

Abstract-Principal curves can learn high-accuracy data from multiple low-accuracy data. However, the current proposed algorithms based on global optimization are too complex and have high computational complexity. To address these problems and in the inspiration of the idea of divide and conquer, this paper proposes a Greedy algorithm based on dichotomy and simple averaging, named as KPCg algorithm. After that, three simulation data sets of sinusoidal, zigzag and spiral trajectories are used to test the performance of the KPCg algorithm and we compare it with the k-segment algorithm proposed by Verbeek. The results show that the KPCg algorithm can efficiently learn high accuracy data from multiple low-accuracy data with constraint endpoints and have advantages in accuracy, computational speed and scope of application.

School/Discipline

Dissertation Note

Provenance

Description

Access Status

Rights

Copyright 2014 Academy Publisher

License

Grant ID

Call number

Persistent link to this record