Master’s studies

Syllabus for Complex Data: Analysis and Visualisation

Komplexa data - analys och visualisering

Syllabus

  • 15 credits
  • Course code: 1MB525
  • Education cycle: Second cycle
  • Main field(s) of study and in-depth level: Technology A1N, Bioinformatics A1N
  • Grading system: Fail (U), 3, 4, 5.
  • Established: 2011-03-07
  • Established by: The Faculty Board of Science and Technology
  • Revised: 2012-04-27
  • Revised by: The Faculty Board of Science and Technology
  • Applies from: week 27, 2012
  • Entry requirements: 120 credits including Scientific Computing II or completed first school year within the Master ' s programmes Molecular biotechnology or Bioinformatics.
  • Responsible department: Biology Education Centre

Learning outcomes

The course intends to give knowledge of analysis and visualisation of complex multidimensional data.
On completion of the course, the student should be able to

  • explain the theoretical basic principles of some model families within the subject area and apply these on biomedical/biological data
  • explain the theoretical basic principles of some general algorithmic modelling technologies within the subject area and apply these on biomedical/biological data
  • independently choose appropriate methods for given biological/biomedical data and issues
  • critically analyse, evaluate and compile biological/biomedical results based on the models that are created from complex amounts of data.

Content

The course treats analysis and visualisation of complex multidimensional data where collected measured values for an observation may consist of a combination of different qualitative and quantitative variables that most often come together with measurement errors (technical variation), sometimes are completely unknown (missing data), and sometimes are censored (due to detection thresholds and/or saturation effects in measuring instruments).
(1) Unsupervised explorative modelling: Compression and visualisation in the form of e.g., factor analysis, clustering, multidimensional scaling, nonnegative matrix factorisation, fitting to manifolds, autoassociative mappings. (2) Supervised modelling: Different families of methods for prediction e.g., model-based, linear, basis expansion based, local, additive, tree based, rule based, neural network based, support vector based, prototype based. (3) General algorithmic modelling techniques such as ensemble methods, bagging, boosting, Bayesian inference, performance evaluation, model selection, variable selection.

Instruction

Lectures, seminars, computer exercises and projects.

Assessment

Theory 5 credits, Implementation and projects 10 credits. For Pass grade, participation in at least 80% of the implementation and project meetings, as well as an approved project report are required. The theory part is examined in a written examination and in compulsory written assignments.

Reading list

Applies from: week 27, 2012

  • Hastie, Trevor; Tibshirani, Robert; Friedman, Jerome The elements of statistical learning : data mining, inference, and prediction

    2. ed.: New York: Springer, 2009

    Find in the library

    Mandatory