Phone: +31 13 466 25 44 (secretary)
Personal webpage Shuai Yuan
Identifying Group Differences in Large-Scale Multi-block Data
The main theme of my project is clustering analysis on multi-source data. Psychological studies more and more often yield multi-source data, which consists of novel blocks of data (e.g. genetic data) and traditional blocks of data (e.g. survey data) collected from the same sample. Such data is (or at least has the potential to be) the trash trove for researches, in that they could not only reveal complex social mechanisms where several influences act together, but moreover offers insights into the differences in such mechanisms between unknown subgroups. Such insights will be invaluable in practical researches. For example, it could suggest effective intervention for a certain target group. The development of valid multi-source clustering method is challenging, however, since the appropriate methods should at least achieve three critical goals: (1) correctly detect the subgroups in the samples, (2) successfully identify the group-specific mechanisms and (3) capable of dealing with potential high-dimensional data. Aiming at tackling these important challenges, in the current project, we will develop and disseminate novel statistical tools that 1) find the different sets of linked variables that underlie complex social phenomena where several influences are at play and 2) predict an outcome based on such diverse sets of linked variables.
Prof. dr. J.K. Vermunt, Dr. K. van Deun, Dr. K. de Roover
NWO Research Talent Grant
1 October 2017-30 September 2021