Home > Courses


Computational Multivariate Statistics (AMCS 309), Fall 2013 - Prof. Marco Scavino

  • Class schedule:  Sunday from 16:00 to 17:30 and Tuesday from 10:30 to 12:00
          Starts on Sunday, September 1st and ends on Tuesday, December 3rd
  • Location: Room 3222, Building 9  
Once simultaneous measurements on many variables are available by virtue of a sampling strategy or an experimental design, multivariate statistical models can be used to extract relevant information to support or reject the research hypotheses that have been raised.
This course introduces some basic multivariate statistical techniques, taking into account the mathematical theory, their computational issues and their rigorous application in order to explore and analyze real-world data sets.
Multiple and multivariate regression methods are considered, focusing on their goodness-of-fit and their predictive ability, as well as on variable selection procedures and their assessment.
Dimension reduction techniques, discrimination techniques and clustering methods are introduced, in order to achieve understanding and a proper interpretation of the data, and to discover relevant groups in structured and non structured data.
The multivariate reduced-rank regression model will provide a conceptual framework for some classical statistical techniques, which were independently proposed to accomplish different goals.
Along the course we will use the R language, a free and open source environment for statistical computing and graphics (http://www.r-project.org).