Project title: Stream mining of content popularity evolution and its applications

Supervisor name: Dr. Saverio Niccolini, Dr. Mohamed Ahmed or someone with the same level of experience and/or expertise.

PhD enrolment: N (this is a postdoc position)

Objectives: Given the dominant role in volume of multi-media content in today’s internet traffic and its expected continued growth, this work will look at how this traffic can be profiled and characterised accurately. The candidates will take part in developing new and innovative algorithms to help explain how users consume content, and predict content popularity growth patterns. The roles will combine both theoretical and practical approaches; candidates will work within a data driven role, and will help to develop highly scalable algorithms and processing platforms. This position will have an algorithmic focus and the candidate will be charged with developing the principles, methods and statistical tools help understand and predict content popularity.

Tasks and methodology: design and evaluation of machine learning techniques able to achieve the objectives specified above, analysis of real-world large data sets is foreseen

Results: the candidate is expected to deliver scalable and accurate algorithms for mining large data sets.

Dissemination: 1 to 2 publications per year to either conferences, magazine or journals

Planned secondment: SamKnows (during the third year for a period from 3 to 6 months). Samknows is an SME that supplies broadband benchmarking services to governments. Thus joining their engineers will provide the postdoc student the sufficient expertise and hand on training measurements. An alternative secondment might also be possible in Telefonica which is one of the major European operators.