Multivariate statistical methods provide a powerful tool for analyzing data when observations are taken over a period of time on the same subject. With the advent of fast and efficient computers and the availability of computer packages such as S–plus and SAS, multivariate methods once too complex to tackle are now within reach of most researchers and data analysts. With an emphasis on computing techniques in combination with a full understanding of the mathematics behind the methods, Methods of Multivariate Statistics offers an up–to–date account of multivariate methods. Focusing on the maximum likelihood method for estimation, testing of hypotheses, and "profile analysis," this book offers comprehensive discussions of commonly encountered multivariate data and also covers some practical and important problems lacking in other texts. These include:
∗ Missing at–random observations
∗ "Growth Curve Models" and multivariate one–sided tests applicable in pharmaceutical and medical trials
∗ Bootstrap methods
∗ Principal component method for predicting a multivariate response vector
∗ Outlier detection and handling inference when covariance is singular
With clear chapter introductions and numerous problem sets, Methods of Multivariate Statistics meets every statistician′s need for a comprehensive investigation of the latest methods in multivariate statistics.
Multivariate Methods: An Overview.
Multivariate Normal Distributions.
Outliers Detection and Normality Check.
Inference on Location Hotelling′s T2.
Multivariate Analysis of Variance.
Classification and Discrimination.
Growth Curve Models.
Principal Component Analysis.
Inference on Covariance Matrices.
Missing Observations: General Case.
Missing Observations: Monotone Sample.
Imputting Missing Data.
Some Results on Matrices.