Probabilities of Correct Classification in Discriminant Analysis

“…Dunn and Varady [11] studied the relationship between pa and P2 for samples with no missing values. Following the line of Dunn and Varady, the present study investigates Pa when the samples have missing values and a method of handling missing values is applied.…”

Section: (22)mentioning

confidence: 99%

The Treatment of Missing Values in Discriminant Analysis—I. The Sampling Experiment

Chan

Dunn

1972

Journal of the American Statistical Association

Self Cite

View full text Add to dashboard Cite

2.1) (2.3) Probabilitie. of correct cla..ificafion under several commonly u.ed method. of handling mi.sing value. are .tudied by Monte Carlo method•. The method. include u'e of only complete observation vectors; u'e of all observation. with no replacement; .ub.titution of mean. for missing ob.ervations; Buck's regression method; and, Dear'. principal component method. Discriminant functions were formed u.ing independent random sample. from two multivariate normal distribution. with equal covariance matrices. Mis.ing values occur randomly in each variable and independently of mi..ing value. in other variable•. The mean .ubstitution method and principal component method are, in general, .uperior to the other methods for ca.e. considered.

show abstract

“…In such situations, unknowns from unanalyzed sources can be incorrectly assigned by discriminant analysis to one of the known source areas in the study (Hess, 1996). Second, the current study involves only 11 samples per source, and both Dunn and Varady (1966) and Luedtke (1979) suggest that 25 samples per source is a minimum required for discriminant analysis. Accordingly, further evaluations of the differences between the Black Canyon and Sedillo Hill source areas are made, in this 2 It has been suggested that some difficulties may arise in using Ga and Pb to distinguish between Black Canyon and Sedillo Hill quarries.…”

Section: Statistical Comparisons Of Black Canyon and Sedillo Hill Silmentioning

confidence: 99%

Geochemical comparisons of silicified rhyolites from two prehistoric quarries and 11 prehistoric projectile points, Socorro County, New Mexico, U.S.A.

Dello‐Russo¹

2004

Geoarchaeology

View full text Add to dashboard Cite

This paper documents the results of nondestructive X-ray fluorescence (XRF) geochemical analyses and comparisons of silicified rhyolites from the Black Canyon and Sedillo Hill prehistoric quarries near Socorro, New Mexico, and subsequent comparisons of the quarry rhyolites with 11 temporally diagnostic projectile points. At the current level of analysis, findings indicate that (1) the two quarries are chemically distinct and (2) the lithic materials of two projectile points match the silicified rhyolite from the Black Canyon quarry, suggesting quarry use during the Early Archaic period (4800-3200 B.C.) and the Late Archaic period (1800-800 B.C.). A close match between the Black Canyon rhyolite and one other point provides tentative support for use of the quarry during the Clovis Paleoindian period (9500-9000 B.C.). Additional geochemical analyses of the two quarries and other regional sources should be undertaken to verify these results and further expand our understanding of prehistoric mobility in the Southwest. ᭧

show abstract

Probabilities of Correct Classification in Discriminant Analysis

Cited by 51 publications

References 4 publications

The expectation of Mahalanobis' generalized distance

The expectation of Mahalanobis' generalized distance

The Treatment of Missing Values in Discriminant Analysis—I. The Sampling Experiment

Geochemical comparisons of silicified rhyolites from two prehistoric quarries and 11 prehistoric projectile points, Socorro County, New Mexico, U.S.A.

Contact Info

Product

Resources

About