Options
Some multivariate techniques for analyzing lizard data
Author
Liew, Kok Keong
Supervisor
Yap, Sook Fwe
Abstract
Researchers in biological, physical and social sciences often collect multivariate measurements in a single study. Although data analysis may be performed on the individual variable, some forms of analyses of these multivariate data require the application of multivariate techniques.
The aim of this Academic Exercise is to explore some of the multivariate techniques we can use for two sets of lizard data provided by Dr C H Diong from Natural Science, National Institute of Education, Singapore. The objectives are: (1) investigation of differences among the means of the variables used, (2) reduction of the dimensionality of the data sets and (3) derivation of rules for classifying lizards by gender or species. The multivariate techniques that are considered here include one-way multivariate analysis of variance (MANOVA), principal component analysis (PCA), discrimination and classification analysis.
For each multivariate technique, the essential elements of the statistical methodology are first presented as understood from the main reference by Johnson and Wichern (2002), together with some of the author's own derivations and interpretations of the theory. Next, the results of applying the multivariate techniques to the two lizard data sets are reported and interpreted in biological context. The study provides a comprehensive account of the theory of the three multivariate techniques and their applications to the analysis of the lizard data.
The aim of this Academic Exercise is to explore some of the multivariate techniques we can use for two sets of lizard data provided by Dr C H Diong from Natural Science, National Institute of Education, Singapore. The objectives are: (1) investigation of differences among the means of the variables used, (2) reduction of the dimensionality of the data sets and (3) derivation of rules for classifying lizards by gender or species. The multivariate techniques that are considered here include one-way multivariate analysis of variance (MANOVA), principal component analysis (PCA), discrimination and classification analysis.
For each multivariate technique, the essential elements of the statistical methodology are first presented as understood from the main reference by Johnson and Wichern (2002), together with some of the author's own derivations and interpretations of the theory. Next, the results of applying the multivariate techniques to the two lizard data sets are reported and interpreted in biological context. The study provides a comprehensive account of the theory of the three multivariate techniques and their applications to the analysis of the lizard data.
Date Issued
2004
Call Number
QA278 Lie
Date Submitted
2004