Data analysis in biomedicine

Facts  
Duration: 1 semester
Period: Spring (2) Semester
Credits: 3 ECTS
Contact Hours: 24
Hours: 108

Main Objectives

The course examines the basic geometric and statistical pattern recognition methods applied to assessing the state of biosystems and making decisions in medicine and covers the basics of submission and publication of the results of statistical analysis of biomedical data.

Learning Outcomes

As a result of the course, a student must:

  • know: the principles of statistical methods of description and analysis of multidimensional biomedical data; geometric and statistical approaches to pattern recognition used in medical diagnosis;
  • be able to: conduct a generalized description of biological systems’ characteristics based on the analysis of primary biomedical data; assess the information content of a Biomedical index in classification issues and decision-making in biological systems;
  • master: the skills of using computer multivariate statistical analysis of biomedical data in the assessment of biosystems issues.

Professor

 Vasily Fokin

Apply

Read more

Course annotation

Data analysis in biomedicine (3 European Credits)

Taught by: Prof. Vasily Fokin

The module discusses basic mathematical approaches to multidimensional biomedical data analysis and construction of mathematical models for biological systems at various organization levels. To take this module students are required to have knowledge of basic concepts of mathematical analysis, vector algebra, probability theory and computer skills. The course begins with a discussion of the features of multidimensional biomedical data statistical description. The course examines the basic geometric and statistical pattern recognition methods applied to assessing the state of biosystems and making decisions in medicine. In conclusion covers the basics of submission and publication of the results of statistical analysis of biomedical data.

The module covers the following topics:

  • Description and analysis of biomedical data.
  • Mathematical methods for decision-making in medicine.
  • Publication of the results of the statistical analysis of biomedical data.

Learning objectives:

As a result of the course, a student must:

  • know: the principles of statistical methods of description and analysis of multidimensional biomedical data; geometric and statistical approaches to pattern recognition used in medical diagnosis;
  • be able to: conduct a generalized description of biological systems’ characteristics based on the analysis of primary biomedical data; assess the information content of a Biomedical index in classification issues and decision-making in biological systems
  • master: the skills of using computer multivariate statistical analysis of biomedical data in the assessment of biosystems issues.

Content of the module:

Capabilities of mathematical methods of data analysis in the study of biological systems. Biomedical data as a source of information about the state of biosystems. Classification of mathematical methods of biomedical data analysis. Information technology analysis. Linear discriminant functions. Evaluation of the results of clustering and quality criteria of recognition. A probabilistic approach to solving classification issues. Classification errors. Matrix loss and estimating the average risk of misclassification. Bayesian classifier. Evaluation of informative content of features. Recognition training. Perceptron model and algorithm. An overview of modern approaches to solving classification issues: methods of potential functions, neural networks, methods of decision rules, formal language theory (linguistic methods). Recommendations and requirements for publication of the results of biomedical research. General algorithm for analysis of scientific medical publications.

Overview of tasks and lectures:

6 two-hour lectures and 6 practical classes totaling 12 hours in the second semester.

Topics of lectures:

1. General characteristics of biomedical data.

2. Preliminary analysis of data.

3. Logic of statistical inference.

4. Qualitative data analysis.

5. Decision-making in medicine (a geometric approach).

6. Decision-making in medicine (a statistical approach). Presentation of the results of the statistical analysis in scientific publications.

Practical classes on topics:

  • Description and multivariate statistical analysis of biomedical data.
  • Decision making in medicine.
  • Presentation of the results of the statistical analysis in scientific publications.

Position within the programme:

This is a unique module examining the application of multidimensional biomedical data mathematical methods of analysis in studying and evaluating of biosystems. The acquired knowledge and skills are relevant to the research methodology of designing properties of biological systems with the use of computer technology.

Teaching format

Structure

The module is scheduled for the second semesters. The course involves 108 hours, including 24 hours in the classroom. The sections of the course are given 6 weeks (6 lectures of 2 hours and 6 practical classes of 2 hours) in the second semester.

Grading

The form of final assessment is an exam. The exam requires satisfactory performance in individual practical studies.