Optimisation and mining of data streams

A data stream can be defined as a flow of instances arriving continuously in an ordered sequence, potentially unbounded in size. Such streams are prevalent today, and it is desirable for methods capable of extracting knowledge from this data to be developed. However, processing data streams presents additional challenges when compared to static data. For example, the dynamic nature of streams means that the underlying distribution can change over time (concept drift), which must be accounted for. The aim of this work will be to investigate and develop techniques for effectively learning from drifting data streams, as well as incorporating optimisation methods in order to generate and update solutions to real-world problems in response to the arriving data.