Network Multiple Station Discriminant Functions

Abstract

An extension is made of linear discriminant analysis to the case where multiple station observations are available for each event. Multivariate regression is used to estimate the mean vectors and covariance matrix in the multiple station discriminant function. The same stations need not be observed for all events as a separate discriminant vector is derived for each station observation and only the discriminant functions available for each event are added. The identification curves for this multiple station discriminant are calculated for surface body wave pairs from a population of 20 combinations of six LRSM stations. The results obtained indicate that the multiple station discriminant function is superior to the usual method which treats the mean vector for each event as a single observation, a result of importance in the application of discriminants by a network of stations.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 03, 1972
Accession Number
AD0758224

Entities

People

  • R. R. Blandford
  • R. Shumway

Organizations

  • Teledyne Technologies

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Computer Programs
  • Covariance
  • Data Science
  • Databases
  • Detection
  • Discriminant Analysis
  • Discrimination
  • Earthquakes
  • Explosions
  • False Alarms
  • Identification
  • Information Science
  • Normal Distribution
  • Observation
  • Probability
  • Seismic Discrimination
  • Surface Waves

Readers

  • Regression Analysis.
  • Seismology