To be able to, for instance, come to conclusions concerning the future shopping behaviour of customers based on statistics is one part of the results in a new dissertation in econometrics at Linnaeus University, written by Deliang Dai.
Deliang Dai's research focuses on analysis of high-dimensional data, that is to say large quantities of data, where each observation includes a large number of variables. The larger the data set to be analysed, the harder it can be to come up with accurate conclusions. Deliang Dai is part of a research group at Linnaeus University that works to develop methods that will make it possible to come to more accurate conclusions based on this type of statistical surveys.
"These methods can, for instance, be used by the online trade to analyse the shopping behaviour of their customers and to some extent predict future purchases by these customers. This will make it possible for the companies to work more with target groups, better plan their stock and so on", explains Dai.
"My continued research will focus on how the accurateness of the results is affected by different conditions and how the method can be modified depending on in what situation it is to be used" Dai continues.
Deliang Dai originally comes from China. He moved to Sweden to study for his master's degree. He came to Linnaeus University in 2012 for his doctoral studies.
"Throughout my doctoral studies, I've also been teaching while doing research and I think that these two parallel processes have been cross-fertilising. I'm inspired by others, both by my students and by my teachers and supervisors. To carry out research is somewhat like walking through a maze. A new reality presents itself behind every corner and one has to be smart to find the best way out", Dai concludes.
Dissertation title: "On high-dimensional Mahalanobis Distances"
Supervisor: Professor Thomas Holgersson, Linnaeus University