(2017 - 2020) Helmholtz Association [total funding of 2.2 Mio. €]
joint project with HMGU, DZNE, DKFZ, FZJ, GFZ/HPI, HZI, MDC, UFZ
Large data sets with many variables frequently contain unobserved, missing or noisy entries. Dealing with these missing values is crucial for any later step of data analytics. Only when properly dealing with these sparse data sets, including the combination of multiple sparse observations of the same entity from different views, we can hope to achieve meaningful big data analytics results. Hence we develop, evaluate and share methods for (1) data imputation in order to fill these missing values with statistical methods and (2) big data analytics making these methods aware of missing values. Our technology will be an enabler for many research areas ranging from patient data in medicine to remote sensing in geography.