Asymptotic Distribution of the Correlation Matrix in a Monotone Incomplete Sample

Abstract

Data with missing values are often obtained using multivariate statistical analyses. It is crucial to study how to estimate parameters and test hypotheses using such data. There exists a step monotone incomplete sample as a simple model of data, which includes such missing values. In this study, we derive the asymptotic distribution of the estimator for the correlation matrix and propose a hypothesis testing method for it in a three-step monotone incomplete sample. Further, we investigate the accuracy of our results by numerical simulation.

Keywords

Asymptotic distribution / Correlation matrix / Monotone incomplete sample / 62E20 / 62F12 / 62H12 / 62H15

References

[1] AndersonT, OlkinI. Maximum likelihood estimation of the parameters of a multivariate normal distribution. Linear Algebra Appl., 1985, 70: 147-171.

[2] AndersonTW. Maximum likelihood estimates for multivariate normal distribution when some observations are missing. J. Am. Stat. Assoc., 1957, 52: 200-203.

[3] ChangWY, RichardsSP. Finite-sample inference with monotone incomplete multivariate normal data, I. J. Multivar. Anal., 2009, 100: 1883-1899.

[4] ChangWY, RichardsSP. Finite-sample inference with monotone incomplete multivariate normal data, II. J. Multivar. Anal., 2010, 101: 603-620.

[5] DempsterA, LairdNM, RubinDB. Maximum likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. Ser. B, 1977, 39: 1-38.

[6] FujisawaH. A note on the maximum likelihood estimators for multivariate normal distribution with monotone data. Commun. Stat. Theory Methods, 1995, 24: 1377-1382.

[7] GuptaAK, NagarDKMatrix Variate Distributions, 2000, Boca Raton. Chapman & Hall.

[8] HaoJ, KrishnamoorthyK. Inferences on a normal covariance matrix and generalized variance with monotone missing data. J. Multivar. Anal., 2001, 78: 62-82.

[9] HyodoM, ShutohN. Asymptotic power comparison of ${T^2}$-type test and likelihood ratio test for a mean vector based on two-step monotone missing data. Commun. Stat. Theory Methods, 2020, 49(17): 4270-4287.

[10] HyodoM, ShutohN, SeoT, PavlenkoT. Estimation of the covariance matrix with two-step monotone missing data. Commun. Stat. Theory Methods, 2016, 45: 1910-1922.

[11] JinadasaKG, TracyDS. Maximum likelihood estimation for multivariate normal distribution with monotone sample. Commun. Stat. Theory Methods, 1992, 21: 41-50.

[12] KandaT, FujikoshiY. Some basic properties of the MLE’s for a multivariate normal distribution with monotone missing data. Am. J. Math. Manag. Sci., 1998, 18: 161-190

[13] KawasakiT, SeoT. Bias correction for ${T^2}$ type statistic with two-step monotone missing data. Statistics, 2016, 50: 76-88.

[14] KawasakiT, ShutohN, SeoT. On the asymptotic distribution of ${T^2}$-type statistic with two-step monotone missing data. J. Stat. Theory Pract., 2018, 12: 657-668.

[15] LittleRJA, RubinDBStatistical Analysis with Missing Data, 20022Hoboken. Wiley-Interscience. .

[16] RichardsDSP, YamadaT. The stein phenomenon for monotone incomplete multivariate normal data. J. Multivar. Anal., 2010, 101: 657-678.

[17] ShutohN, KawasakiT, SeoT. Testing equality of mean vectors in two sample problem with missing data. Commun. Stat. Simul. Comput., 2010, 39: 487-500.

[18] SrivastavaMS. Multivariate data with missing observations. Commun. Stat. Theory Methods, 1985, 14: 775-792.

[19] TsukadaS. Asymptotic expansion for distribution of the trace of a covariance matrix under a two-step monotone incomplete sample. J. Multivar. Anal., 2014, 129: 206-219.

[20] TsukadaS. Equivalence testing of mean vector and covariance matrix for multi-populations under a two-step monotone incomplete sample. J. Multivar. Anal., 2014, 132: 183-196.

[21] TsukadaS. Unbiased estimator for a covariance matrix under two-step monotone incomplete sample. Commun. Stat. Theory Methods, 2014, 43: 1613-1629.

[22] TsukadaS. Asymptotic properties of a correlation matrix under a two-step monotone incomplete sample. Linear Algebra Appl., 2016, 488: 86-101.

[23] TsukadaS. Unbiased estimator for a covariance matrix in a three-step monotone incomplete sample. Linear Algebra Appl., 2019, 579: 169-205.

[24] van BuurenSFlexible Imputation of Missing Data, 2012, Boca Raton. Chapman & Hall. .

[25] YagiA, SeoT. A test for mean vector and simultaneous confidence intervals with three-step monotone missing data. Am. J. Math. Manag. Sci., 2014, 33: 161-175

[26] YagiA, SeoT. Tests for equality of mean vectors and simultaneous confidence intervals with two-step or three-step monotone missing data patterns. Am. J. Math. Manag. Sci., 2015, 34: 213-233

RIGHTS & PERMISSIONS

School of Mathematical Sciences, University of Science and Technology of China and Springer-Verlag GmbH Germany, part of Springer Nature