Degree | Type | Year | Semester |
---|---|---|---|
2503852 Applied Statistics | OB | 2 | 2 |
A previous course of Linear Algebra is essential, as well as courses in Probability and Statistical Inference. Also, a good knowledge of the R software is assumed.
Most of collected data sets are multivariate, that is, for the same experimental unit, perhaps a complex nature object, we observe simultaneously the values of several variables. Multivariate Analysis deals with the methods that are most appropriate for describing, exploring and modelling vector data, as well as for applying statistical inference. The interest in processing large amounts of observations in many variables of a diverse nature, together with the aim of reducing the information that is not relevant or discovering patterns of association between variables or between cases, they have recently promoted the development of a series of multivariate techniques. This subject is intended as a first contact of the student with the statistical learning theory. Students must understand the power and applicability as well as the limitations of the multivariate tools, some of which are based on very simple heuristic ideas. The subject focuses in the applications, mostly in the computer work sessions using the R free software resources. Theoretical and problems sessions are devoted to formalize the models, derive their properties, and study some models validation techniques.
Statistical learning and dimension reduction
Factorial methods I: Principal components analysis (PCA)
Factorial methods II: Factorial analysis (FA)
Factorial methods III: Multidimensional scaling (MDS) and correspondence analysis (CA)
Cluster analysis (CLA)
Multivariate inference basics
Discriminant analysis (DA) and other supervised methods
The theoretical sessions, where the multivariate methods will be exposed in detail and discussed on the bases of appropriate examples. The classroom presentations will be posted on the virtual campus. The revision and expansion of contents using the course bibliography will be encouraged.
The computer lab sessions are designed to be implemented in statistical software R. The exercises statements and other auxiliary material will be made available to the students in the Virtual Campus. Extension exercises will be proposed to be solved autonomously.
The theoretical sessions, where the multivariate methods will be exposed in detail and discussed on the bases of appropriate examples. The classroom presentations will be posted on the virtual campus. The revision and expansion of contents using the course bibliography will be encouraged.
The computer lab sessions are designed to be implemented in statistical software R. The exercises statements and other auxiliary material will be made available to the students in the Virtual Campus. Extension exercises will be proposed to be solved autonomously.
The collaboration and participation of all students will be sought, without discrimination based on sex or any other cause.
Title | Hours | ECTS | Learning Outcomes |
---|---|---|---|
Type: Directed | |||
Computer lab sessions | 26 | 1.04 | 1, 4, 14, 5, 6, 8, 10, 11, 15, 7 |
Theoretical classes | 26 | 1.04 | 1, 9, 3, 2, 4, 14, 5, 6, 8 |
Type: Autonomous | |||
Personal work | 42 | 1.68 | 9, 4, 5, 6, 13, 7 |
Tasks solving and delivery | 44 | 1.76 | 1, 2, 4, 14, 8, 13, 12, 10, 11, 15, 7 |
The course grade (NC) will be calculated on the basis of the delivered tasks and the marks in two partial exams (P1 and P2), including both theoretical and computational exercices:
NC = 0.4• P1 + 0.5 • P2 + 0.10 • Lli
where P1 and P2 correspond to the first and second partial grades, respectively, and Lli is based on the delivered tasks and will not be recoverable.
In order to succeed in this course, it is mandatory that NC>=5 and P1>3.5 and P2>3.5. Besides that, the students will have the option of taking an additional recovery exam (F) with the same format (theoretical and computational questions). The final qualification will be:
NF = Max (NC, 0.90 • F + 0.10• Lli)
Observation: Only students who have participated in 2/3 of the continuous assessment activities will have the recovery option. Honor grades will be granted at the first complete evaluation. Once given, they will no be withdrawn even if another student obtains a larger grade after consideration of the final exam.
Title | Weighting | Hours | ECTS | Learning Outcomes |
---|---|---|---|---|
Partial exam 1 | 0,4 | 5 | 0.2 | 1, 14, 5, 6, 13, 12 |
Partial exam 2 | 0,5 | 5 | 0.2 | 1, 2, 4, 5, 6, 8, 13, 12, 10, 7 |
Tasks delivery | 0,1 | 2 | 0.08 | 9, 3, 8, 13, 12, 10, 11, 15, 7 |
Everitt, B., Hothorn, T. ; An introduction to Applied Multivariate Analysis with R. Springer, 2011.
Härdle, W., Simar, L.; Applied Multivariate Statistical Analysis. Springer,2007.
Peña, D.; Análisis de datos multivariantes. McGraw Hill, 2002.
Rencher, A., Christensen, W.; Methods of Multivariate Analysis. Wiley Series in Probability and Mathematical Statistics, 2012.
Complementary references
Coghlan, A.; Little book of R for Multivariate Analysis.
https://little-book-of-r-for-multivariate-analysis.readthedocs.io/en/latest/
Cuadras, C.; Nuevos Métodos de Análisis Multivariante (web), 2014.
Greenacre, M.; La pràctica del análisis de correspondencias. Fundacion BBA, 2003.
James, G., Witten, D., Hastie, T., Tibshirani, R.; An Introduction to Statistical Learning. Springer, 2014.
Mardia, K.V, Kent, J.T., Bibby, J.M.; Multivariate Analysis. Academic Press, 2003.
Rencher, A.; Multivariate Statistical Inference and Applications. John Wiley &Sons, 1998.