TILM3540 Introduction to multivariate methods
This course has a moodle2 page
Objectives:
A practical introduction to classical analysis of multivariate data when the response variable is multivariate and continuous. Basic understanding of the mathematical background and use of the traditional multivariate analysis tools.
Outline:
- Description of multivariate data: data matrix, mean vector, covariance matrix, other descriptive statistics, Mahalanobis distance, scatter plots
- Multivariate normal distribution and its properties
- Multivariate location tests: Hotelling’s T2, MANOVA and multivariate linear regression
- The use of covariance matrix: Principal component analysis (PA), canonical correlation (CA)
- Factor analysis (FA) and independent component analysis (ICA)
- Discriminant analysis and cluster analysis
- Target group: Student is statistics (A) and doctoral students in other disciplines
- Prerequisites: A good knowledge of univariate statistical methods
- Teaching methods: Lectures (20 hours) and practicals in pc class (10 hours)
- Lectures on Mondays at 10-14, Pub3, Publicum, 2.9.2013 – 16.9.201
- Lectures on Thursdays at 12-16, Pub2, Publicum, 5.9.2013 – 12.9.2013
- 10 hours of practicals, two per week, detailed schedule will be updated during August. There will be multiple groups depending on the number of participants.
- Evaluation: Active participation in practicals and an essay (other disciplines: passed/failed) or active participation and an exam (students in statistics, 0-5).