ST504: Prediction and Classification (5 ECTS)

STADS: 25000401

Level
Bachelor course

Teaching period
The course is offered in the spring semester.
3rd quarter.

Teacher responsible
No responsible teachers found, contact the department if necessary

Timetable
Show entire timetable
Show personal time table for this course.

Revison of timetable:
: Forelæsning mandag flyttet til kl. 08-10 efter ønske fra de studerende og IMADA.

Prerequisites:
None

Academic preconditions:
One of the courses ST502 Statistical Modelling or ST513 Applied Statistics are recommended.

Course introduction
To gain knowledge about the fundamental ideas and methods for constructing and applying predictors and classificators. The participants will learn to implement and evaluate the methods using the statistical software R.

Qualifications
The participants will achieve competences in:
• understanding when a scientific or practical problem is a prediction or classification problem,
• applying the classical regression model to performing predictions,
• applying the logistic regression model to doing classifications,
• evaluating the quality of a given prediction or classification,
• evaluating the quality of a predictor or classificator,
• evaluating the quality of method for constructing prediktors and classificators,
• constructing efficient predictors and classificators,
• considering the n >> p problem.

Expected learning outcome
After having followed the course the student should be able to:

• understand when a scientific or practical problem is a prediction- or classification problem
• apply classical regression models to predict values of new, future observations
• apply the logistic regression model to perform a classification
• evaluate the quality of a given prediction or classification
• evaluate the quality of a predictor or a classifier
• evaluate the quality of methods to construct predictors or classifiers
• construct effective predictors and classifiers
• apply modern methods for prediction and classification, for example neural network and CART
• consider the n>>p problem

Subject overview
Prediction based on the classical regression model, prediction error, logistic regression, sensitivity and specificity, ROC curves, cross-validation, Akaike’s criterion, variable selection, ridge regression, R-square measures, partial least squares.

Literature
There isn't any litterature for the course at the moment.

Syllabus
See syllabus.

Website
This course uses e-learn (blackboard).

Prerequisites for participating in the exam
None

Assessment and marking:
Oral exam. Internal examiner. Pass/fail.

Project assignment. Internal examiner. Pass/fail.

2 project assignments. The project assignments can be written in groups of max. 3 persons.

Reexamination after 4. quarter.

Expected working hours
The teaching method is based on three phase model.

Forelæsninger (28 timer), eksaminatorier (12 timer), laboratorieøvelser (7 timer).
Educational activities

Language
This course is taught in English, if international students participate. Otherwise the course is taught in Danish.

Course enrollment
See deadline of enrolment.

Tuition fees for single courses
See fees for single courses.