Feature selection and statistical pattern recognition for the classification of ricker wavelets

Kou-Yuan Huang, Weng Yu Shyu, King Sun Fu

Research output: Contribution to conferencePaperpeer-review

1 Scopus citations

Abstract

Feature selection and pattern recognition techniques are used for classification of Ricker wavelets. Envelope and instantaneous frequency are used as the features in the classification procedure. Three feature selection techniques are used. In the information-theoretic approach, divergence and Bhattacharyya distances are used as the criteria of feature selection. Feature space transformation of the discriminant method [maximum of tr(S2-1 S1)] is used to select the optima] feature. From the discriminant method, the optimal eigenvector is selected as the optimal feature. From the experiment, the result shows that instantaneous frequency is chosen as the feature to separate 20 and 30 Hz Ricker wavelets. Using instantaneous frequency as the feature, the Bayes classification result of 20 and 30 Hz Ricker wavelets is quite good.

Original languageEnglish
Pages539-540
Number of pages2
StatePublished - 1 Jan 1984
Event1984 Society of Exploration Geophysicists Annual Meeting, SEG 1984 - Atlanta, United States
Duration: 2 Dec 19846 Dec 1984

Conference

Conference1984 Society of Exploration Geophysicists Annual Meeting, SEG 1984
CountryUnited States
CityAtlanta
Period2/12/846/12/84

Fingerprint Dive into the research topics of 'Feature selection and statistical pattern recognition for the classification of ricker wavelets'. Together they form a unique fingerprint.

Cite this