Time series prediction method based on LS-SVR with modified Gaussian RBF

Yangming Guo; Xiaolei Li; Guanghan Bai; Jiezhong Ma

doi:10.1007/978-3-642-34481-7_2

Time series prediction method based on LS-SVR with modified Gaussian RBF

Yangming Guo, Xiaolei Li, Guanghan Bai, Jiezhong Ma

School of Cybersecurity

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

22 Scopus citations

Abstract

LS-SVR is widely used in time series prediction. For LS-SVR, the selection of appropriate kernel function is a key issue, which has a great impact with the prediction accuracy. Compared with some other feasible kernel functions, Gaussian RBF is always selected as kernel function due to its good features. As a distance functions-based kernel function, Gaussian RBF also has some drawbacks. In this paper, we modified the standard Gaussian RBF to satisfy the two requirements of distance functions-based kernel functions which are fast damping at the place adjacent to the test point and keeping a moderate damping at infinity. The simulation results indicate preliminarily that the modified Gaussian RBF has better performance and can improve the prediction accuracy with LS-SVR.

Original language	English
Title of host publication	Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings
Pages	9-17
Number of pages	9
Edition	PART 2
DOIs	https://doi.org/10.1007/978-3-642-34481-7_2
State	Published - 2012
Event	19th International Conference on Neural Information Processing, ICONIP 2012 - Doha, Qatar Duration: 12 Nov 2012 → 15 Nov 2012

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	7664 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	19th International Conference on Neural Information Processing, ICONIP 2012
Country/Territory	Qatar
City	Doha
Period	12/11/12 → 15/11/12

Keywords

Gaussian RBF
Least squares support vector regression (LS-SVR)
Time series prediction

Access to Document

10.1007/978-3-642-34481-7_2

Cite this

Guo, Y., Li, X., Bai, G., & Ma, J. (2012). Time series prediction method based on LS-SVR with modified Gaussian RBF. In Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings (PART 2 ed., pp. 9-17). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7664 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-34481-7_2

Guo, Yangming ; Li, Xiaolei ; Bai, Guanghan et al. / Time series prediction method based on LS-SVR with modified Gaussian RBF. Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings. PART 2. ed. 2012. pp. 9-17 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{968b16fc8737459b9992789e91cdc50d,

title = "Time series prediction method based on LS-SVR with modified Gaussian RBF",

abstract = "LS-SVR is widely used in time series prediction. For LS-SVR, the selection of appropriate kernel function is a key issue, which has a great impact with the prediction accuracy. Compared with some other feasible kernel functions, Gaussian RBF is always selected as kernel function due to its good features. As a distance functions-based kernel function, Gaussian RBF also has some drawbacks. In this paper, we modified the standard Gaussian RBF to satisfy the two requirements of distance functions-based kernel functions which are fast damping at the place adjacent to the test point and keeping a moderate damping at infinity. The simulation results indicate preliminarily that the modified Gaussian RBF has better performance and can improve the prediction accuracy with LS-SVR.",

keywords = "Gaussian RBF, Least squares support vector regression (LS-SVR), Time series prediction",

author = "Yangming Guo and Xiaolei Li and Guanghan Bai and Jiezhong Ma",

year = "2012",

doi = "10.1007/978-3-642-34481-7_2",

language = "英语",

isbn = "9783642344800",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 2",

pages = "9--17",

booktitle = "Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings",

edition = "PART 2",

note = "19th International Conference on Neural Information Processing, ICONIP 2012 ; Conference date: 12-11-2012 Through 15-11-2012",

}

Guo, Y, Li, X, Bai, G & Ma, J 2012, Time series prediction method based on LS-SVR with modified Gaussian RBF. in Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 7664 LNCS, pp. 9-17, 19th International Conference on Neural Information Processing, ICONIP 2012, Doha, Qatar, 12/11/12. https://doi.org/10.1007/978-3-642-34481-7_2

Time series prediction method based on LS-SVR with modified Gaussian RBF. / Guo, Yangming; Li, Xiaolei; Bai, Guanghan et al.
Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings. PART 2. ed. 2012. p. 9-17 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7664 LNCS, No. PART 2).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Time series prediction method based on LS-SVR with modified Gaussian RBF

AU - Guo, Yangming

AU - Li, Xiaolei

AU - Bai, Guanghan

AU - Ma, Jiezhong

PY - 2012

Y1 - 2012

N2 - LS-SVR is widely used in time series prediction. For LS-SVR, the selection of appropriate kernel function is a key issue, which has a great impact with the prediction accuracy. Compared with some other feasible kernel functions, Gaussian RBF is always selected as kernel function due to its good features. As a distance functions-based kernel function, Gaussian RBF also has some drawbacks. In this paper, we modified the standard Gaussian RBF to satisfy the two requirements of distance functions-based kernel functions which are fast damping at the place adjacent to the test point and keeping a moderate damping at infinity. The simulation results indicate preliminarily that the modified Gaussian RBF has better performance and can improve the prediction accuracy with LS-SVR.

AB - LS-SVR is widely used in time series prediction. For LS-SVR, the selection of appropriate kernel function is a key issue, which has a great impact with the prediction accuracy. Compared with some other feasible kernel functions, Gaussian RBF is always selected as kernel function due to its good features. As a distance functions-based kernel function, Gaussian RBF also has some drawbacks. In this paper, we modified the standard Gaussian RBF to satisfy the two requirements of distance functions-based kernel functions which are fast damping at the place adjacent to the test point and keeping a moderate damping at infinity. The simulation results indicate preliminarily that the modified Gaussian RBF has better performance and can improve the prediction accuracy with LS-SVR.

KW - Gaussian RBF

KW - Least squares support vector regression (LS-SVR)

KW - Time series prediction

UR - http://www.scopus.com/inward/record.url?scp=84869011767&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-34481-7_2

DO - 10.1007/978-3-642-34481-7_2

M3 - 会议稿件

AN - SCOPUS:84869011767

SN - 9783642344800

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 9

EP - 17

BT - Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings

T2 - 19th International Conference on Neural Information Processing, ICONIP 2012

Y2 - 12 November 2012 through 15 November 2012

ER -

Guo Y, Li X, Bai G, Ma J. Time series prediction method based on LS-SVR with modified Gaussian RBF. In Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings. PART 2 ed. 2012. p. 9-17. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-34481-7_2

Time series prediction method based on LS-SVR with modified Gaussian RBF

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this