ECS607U Data Mining

Module code: ECS607U

Credits: 15
Semester: SEM1

The aim of this module is to provide a practical understanding of the use of data mining for decision making. The module will cover the classes of algorithms and models and how an appropriate algorithm is chosen for a given task. The evaluation and interpretation of results and the limitation of methods will be emphasized.

Topics covered are likely to include:

• a taxonomy of problems: classification, pattern recognition and association and of learning types: supervised versus unsupervised; clustering algorithms
• knowledge representations such as regression, decision trees, association rules, bayesian networks
• evaluation techniques: cross validation, performance metrics
• social impact, ethics and the law.

•    * Note that this module is dual level, i.e. it is taught at both levels 6 (year 3) and 7 (year 4). The assessment for the level 6 and 7 variants differs by at least 1/3, either in coursework or exam components, with the higher level variant testing the more advanced learning objectives noted in the relevant module descriptor. Any student who has already studied the level 6 variant may not subsequently study the equivalent level 7 variant.

Level: 6