The Research Foundation – Flanders (FWO) announced that Jefrey Lijffijt will receive a Marie Sklodowska-Curie Fellowship grant, co-funded by the European Research Council, to continue his work in our team for another three years.
The title of his project is Personalised, interactive, and visual exploratory mining of patterns in complex data, and a summary is given below.
Due to technological progress, data collection and storage has grown by several orders of magnitude in volume, and has become a pervasive endeavour in business, science, and other domains. The widespread idea is that this data can be used by scientists, journalists, and businesses to derive new knowledge about reality, innovate, and gain a competitive edge.
Exploratory pattern mining is the endeavour of finding interesting and surprising patterns from large collections of data. There are software tools that help analysts conduct such pattern mining on various kinds of data. However, pattern mining in complex data is currently a tedious task: it requires expertise in statistics, programming skills, and knowledge of specific syntax of software tools. Hence, this kind of analysis is currently done exclusively by Data Scientists, who are educated specifically for this job, and even then it is very labour intensive and time consuming.
We aim to remedy this situation in three ways: (1) guidance of the ‘analyst’ by the computer, by steering the analyst towards patterns that are interesting specifically for him/her, (2) maximally informative visualisation of the patterns, such that the analyst can more quickly internalise the presented information, (3) fluid interaction with the visualisation, in order to enable intuitive feedback into the system about what the analyst is interested in. Altogether, these developments will also enable non-specialists to accessibly explore data.