Abstract:
|
Inferences from multiple data sources can often be fused together to yield to a more effective overall inference than individual sources alone. Such effective fusion learning is of vital importance, especially in light of the automated data acquisition nowadays in many domains. Decision-making processes in many domains such as medicine, life science, social studies, etc. often benefit greatly from considering data from different sources, possibly with varying forms of complexity and heterogeneity in their data structure. This talk presents some newly developed fusion methodologies for extracting and merging useful information. Some methodologies are motivated by challenges arising from massive complex structures from different data sources, while others by specific goal-directed applications, such as in precision medicine. Underlying those methodologies is the concept of "confidence distribution", which, simply put, is a versatile distributional inferential scheme (unlike the usual point or interval inferences) without priors. Some simulation and real applications are also presented.
|