A New Spectrum Estimation Method in Unevenly Sampling Space

Author

Xian, Jun ; Wu, Shuan-Hu ; Liew, Alan ; Smith, David ; Yan, Hong

Author_Institution

Dept. of Electron. Eng., City Univ. of Hong Kong

fYear

2006

fDate

13-16 Aug. 2006

Firstpage

4273

Lastpage

4277

Abstract

Spectrum estimation is a popular method for identifying periodically expressed genes in microarray time series analysis. For unevenly sampled data, a common technique is applying the Lomb-Scargle algorithm. The performance of this method suffers from the effect of noise in the data. In this paper, we propose a new spectrum estimation algorithm for unevenly sampled data. The new method is based on signal reconstructing technic in aliased shift-invariant signal spaces and a direct spectrum estimation formula was derived based on B-spline basis. The new algorithm is very flexible and can reduce the effect of noise by adjusting the order of B-spline basis. The test on simulated noisy signal and typical periodically expressed gene data shows our algorithm is accurate compared with Lomb-Scargle algorithm

Keywords

biology computing; estimation theory; genetics; molecular biophysics; signal reconstruction; signal sampling; splines (mathematics); time series; B-spline basis; Lomb-Scargle algorithm; direct spectrum estimation formula; microarray time series analysis; shift-invariant signal spaces; signal reconstruction; unevenly sampled data; Biochemistry; Computer science; Cybernetics; Machine learning; Machine learning algorithms; Noise reduction; Sampling methods; Signal processing algorithms; Signal reconstruction; Space technology; Spectral analysis; Spline; Testing; B-spline; Lomb-Scargle algorithm; Spectrum estimation; eriodically expressed gene; signal reconstruction; unevenly sampled data;

fLanguage

English

Publisher

ieee

Conference_Titel

Machine Learning and Cybernetics, 2006 International Conference on

Conference_Location

Dalian, China

Print_ISBN

1-4244-0061-9

Type

conf

DOI

10.1109/ICMLC.2006.259011

Filename

4028823