Course detail

Advanced Methods in Biostatistics

FEKT-MPC-STAAcad. year: 2019/2020

The course is designed as a practice-oriented course focused on advanced application of multivariate statistics and stochastic modeling of the biological and medical data. The course follows on the basic methodology of one-dimensional data analysis. The methods of descriptive multivariate analysis with special emphasis on the visibility of graphical multidimensional data, stochastic modeling and prediction are discussed. Theoretical aspects are always given by way of examples and the emphasis is on the practical aspects of teaching. All computational techniques are practiced using commercially available software tools (Statistica for Windows, SPSS).

Language of instruction

Czech

Number of ECTS credits

5

Mode of study

Not applicable.

Guarantor

doc. Ing. Daniel Schwarz, Ph.D.

Department

Department of Biomedical Engineering (UBMI)

Learning outcomes of the course unit

After completion of the course participants will be able to:
• evaluate the assumptions of multivariate data analysis / modeling and select the appropriate method for solving a given problem,
• apply the ordination methods and cluster analysis,
• use the tools of multivariate linear and logistic regression,
• select and use generalized linear models,
• use of multivariate analysis and models in statistical software.

Prerequisites

The subject knowledge of biostatistics (ASTA) on the Bachelor´s degree level is requested, work with PC, work with software Statistica.

Co-requisites

Not applicable.

Planned learning activities and teaching methods

Teaching methods depend on the type of course unit as specified in the article 7 of BUT Rules for Studies and Examinations.

Assesment methods and criteria linked to learning outcomes

Compliance with the requirements for terminating exercise: participation in seminars, 3 tests in more than 50% points.
Final test for more than 50% points. The test is aimed at testing the overview of multivariate statistics and stochastic modeling.

Course curriculum

1. Goals and objectives of multivariate data analysis and modeling, the relationship univariate and multivariate statistical methods
2. Multivariate statistical distribution and testing, operations with vectors and matrices
3. Similarities and distances in a multidimensional space, association matrix I
4. Similarities and distances in a multidimensional space, association matrix II
5. Cluster analysis
6. Ordination analysis - principles of dimensionality reduction
7. Ordination analysis - an overview of methods
8. Discriminant analysis
9.The principles of stochastic modeling
10. Linear models - the basics
11. Logistic regression analysis of ROC curves; Advanced prediction methods - an overview
12. Strategy multivariate analysis of clinical data, multi-dimensional data in clinical trials, bases of meta-analysis
13. Review of methods of time series analysis

Work placements

Not applicable.

Aims

The aim of the course is to provide students skills in multivariate analysis and stochastic modeling of the biological and clinical data.

Specification of controlled education, way of implementation and compensation for absences

Participation in seminar is mandatory, two absences are permitted. In the case of multiple absences, it is possible to substitute the seminar after the agreement with teacher (ideally in another parallel group).

Recommended optional programme components

Not applicable.

Prerequisites and corequisites

Not applicable.

Basic literature

LEGENDRE, Piere a Louis LEGENDRE. Numerical Ecology 2. vyd. Elsevier Science, 1998. ISBN 978-0444892508. (EN)
MELOUN, Milan a Jiří MILITKÝ. Statistické zpracování experimentálních dat. 1. vyd. Praha: Plus, 1994. ISBN 80-85297-56-6. (CS)

Recommended reading

Not applicable.

Classification of course in study plans

  • Programme MPC-BTB Master's 1 year of study, winter semester, compulsory-optional

  • Programme EEKR-CZV lifelong learning

    branch EE-FLE , 1 year of study, winter semester, compulsory-optional

Type of course unit

 

Lecture

26 hod., optionally

Teacher / Lecturer

doc. Ing. Daniel Schwarz, Ph.D.

Syllabus

1. Smysl a cíle vícerozměrné analýzy dat a modelování, vztah jednorozměrných a vícerozměrných statistických metod
2. Vícerozměrné statistické rozdělení a testy, operace s vektory a maticemi
3. Zviditelnění vícerozměrných dat
4. Podobnosti a vzdálenosti ve vícerozměrném prostoru, asociační matice
5. Shluková analýza
6. Ordinační analýzy - principy redukce dimenzionality
7. Ordinační analýzy - přehled metod
8. Diskriminační analýza
9. Principy stochastického modelování
10. Logistická regrese, analýza ROC křivek
11. Zobecněné lineární modely - základy
12. Pokročilé metody predikce - přehled
13. Strategie analýzy vícerozměrných klinických dat, vícerozměrná data v klinických studiích, základy metaanalýzy

Exercise in computer lab

26 hod., compulsory

Teacher / Lecturer

doc. Ing. Daniel Schwarz, Ph.D.

Syllabus

BLOK A: Vizualizace a příprava analýz vícerozměrných dat
1. Metody vizualizace vícerozměrných dat
2. Transformace dat

BLOK B: Aplikace vícerozměrných metod analýzy dat
3. Výpočet asociační matice
4. Hierarchická shluková analýza
5. Nehierarchická shluková analýza
6. Analýza hlavních komponent, statorová analýza
7. Základy metaanalýzy, vícerozměrné vážení, třídění parametrů
8. Diskriminační analýza

BLOK C. Stochastické modelování
9. Lineární regrese
10. Logistická regrese
11. ROC křivky
12. Zobecněné lineární modely
13. Modelová analýza vícerozměrného souboru