Multivariate statistical analysis

Author: He Xiaoqun
Publisher:
Publish Date: 2004-04-01
Features: Multivariate statistical analysis is a very important branch of statistics. In foreign countries, it has been widely applied in natural sciences, management sciences, and social, economic, and other fields since the 1930s. In China, the application of multivariate statistical analysis was initiated in many fields since the 1980s. This book was compiled to meet the new demand. The guiding principle of this book is: while maintaining rigor, it differs significantly from pure mathematical textbooks, emphasizing the application of practical cases and the integration of statistical thinking. It systematically introduces practical methods of multivariate analysis combined with statistical software. To implement this idea, the book references a large number of domestic and international books and literature. While systematically introducing the basic theories and methods of multivariate analysis, it also integrates research examples from social, economic, and natural science fields to combine multivariate analysis methods with practical applications. It pays attention to the close integration of qualitative and quantitative analysis, and strives to incorporate the experience and insights of peers and our own practical application of multivariate analysis. Almost every method emphasizes its respective advantages, disadvantages, and issues to be noted in practical application. To help readers master the content of this book, and considering the applied and practical nature of this course, simple thinking questions and exercises are provided at the end of each chapter. We encourage readers to implement these methods using real-world data. The application of multivariate analysis cannot be separated from computers. The cases in this book mainly use SPSS software, which is widely popular in China, to implement, and some methods are completed using SAS software. A significant feature of this book is that it briefly introduces the practical operation of SPSS or SAS software after each method with examples. References are also provided at the end of each chapter for readers who are interested in further reading. The book is divided into 14 chapters. The main content includes common mainstream methods such as multivariate normal distribution, tests for mean vectors and covariance matrices, cluster analysis, discriminant analysis, principal component analysis, factor analysis, correspondence analysis, and canonical correlation analysis. It also systematically introduces some relatively new methods widely used in fields such as market research, customer satisfaction research, financial research, and environmental research, based on references from domestic and international literature. These include modeling and analysis of qualitative data, log-linear models, BGIS regression, path analysis, structural equation models, conjoint analysis, multivariate graphical representation, and multidimensional scaling. This book can serve as a textbook for undergraduate statistics majors in multivariate analysis courses. Due to the extensive content, teachers can flexibly select topics when using this book as a textbook. It can also be used as a quantitative analysis textbook for graduate students in non-statistics majors. Based on our many years of teaching experience, a 72-hour lecture is most suitable for this book. With the support of computers and projection equipment, teaching will be more convenient and effective.