When data are incomplete, missing data must be handled either in the statistical analysis itself or prior to carrying out statistical analyses. Many techniques for handling missing data exist. However, not all of these techniques can be used for all multivariate analysis problems. Moreover, much remains unknown about how well these methods work under which circumstances for different multivariate analyses techniques.

 

Project goals

1. to develop new techniques for handling missing data in multivariate analysis;

2. to improve existing techniques for handling missing data so that they are more widely applicable to different multivariate analysis techniques;

3. to test these methods under different circumstances and for different multivariate techniques by means of simulations and real data examples.