Admission requirements
Admission only after intake, please see our website
Description
The course will provide an overview of data mining and data stream mining techniques and their applications to cyber security. The course teaches you how to go from raw input data to useful results in credit-card fraud detection, real-time detection of attacks and deviations in SCADA systems, and detection of botnet activities in network traffic. Practical work with real data will provide the participants some hands-on experience with applying data mining to real-life problems. Limitations of data mining in cyber security will also be discussed, in particular the ability of adversaries to modify their data and how to deal with this. Privacy issues and solutions to increase the privacy of individuals in a data set will be reviewed.
Course objectives
Participants will get:
Knowledge and understanding of the potentials and pitfalls of using data mining in cyber security
knowledge and understanding of the inner workings of data mining solutions for cyber data * mining data where the majority of data are benign * detecting anomalies * mining sequential data * learning profiles and fingerprints
the ability to correctly apply data mining tools on real-world problems
knowledge and understanding of privacy and adversarial aware data mining
Prerequisite knowledge is a basic understanding of statistics. Experience with machine learning or data mining algorithms is useful but not required.
Timetable
On the right-hand side of the programme front page of the E-Prospectus you will find a link to the online timetables.
Mode of instruction
Lectures, exercises, class discussion
Lecturers: Dr. Sicco Verwer (TUD)
Assessment method
Lab 1
40% of final grade
Grade must be 5.50 or higher to pass the course
Resit of a fail is possible.
Resit will take the same form
Lab 2
30% of final grade
Grade must be compensated
Resit not possible
Lab 3
30% of final grade
Grade must be compensated
Resit not possible
Only assessments with the weight of 30% and lower are compensable. This means that one does not have to pass an assessment if it weighs 30% or less in order to pass the course, if the average of all assessments combined is at least a 5.5. In addition, assignments weighing up to and including 30% are not re-sitable, meaning that if one failed an assessment of 30% or less one is not allowed to redo it and that assessment must be compensated by the other assessment(s).
Reading list
Compulsory literature and literature for further consultation will be announced via Brightspace
Registration
No registration is required for lectures and exams.
Contact
Dr. Sicco Verwer M.I. Warner, study coordinator
Remarks
For more information see the website see our website