Workshop on Statistical Issues in Analyzing Information from Diverse Sources

May 6 - 7, 2010
CICCADA Center, CoRE Building, Rutgers University

Organizers:

Tami Carpenter, Rutgers University, tcar at dimacs.rutgers.edu

Minge Xie, Rutgers University, mxie at stat.rutgers.edu

Announcement

Today's decision makers in fields ranging from engineering to medicine to homeland security have available to them remarkable new technologies for gathering potentially huge amounts of information from diverse sources. Yet, decision makers are often at a loss as to how to effectively combine and analyze information from disparate sources. Instead they often look at different sources separately and invoke relatively ad hoc methods for gaining combined intuition. This approach may lack rigor and fail to harness the full value of the information at hand.

This workshop will explore methods of, theory for, and barriers to combining data from multiple sources for improved decision making that exploits inferences that are typically more efficient and potentially more accurate than those from any single source. The workshop will bring together statisticians, applied mathematicians, computer scientists and policy makers to address issues related to combining information, and it will disseminate research results in the areas of model building, Bayesian analysis, incorporation of expert opinion, meta-analysis, and machine learning, among others. The topics will encompass statistical and mathematical approaches as well as computational tools that are related to combining information from different sources for inference, learning and decision making.

Questions of potential interest during the workshop include:

• How can we combine information from different studies of the same type?
• To what extent can we combine information from the different types of studies and how do we do it? Can we correct data quality problems across multiple data sources?
• Can use multiple data sources for improved reasoning with uncertain data?
• How do we combine expert opinions or intelligence information (possibly using machine learning approaches) with more traditional quantitative data?
• Should we integrate databases from different studies or should we make combined inferences over different studies?
• How can we combine multiple views to rank items in recommendation systems?
• Can we assure privacy with little or no negative impact on the quality of the results?

The workshop will examine timely and important applications from a variety of fields. These include medicine and public health, as well as applications from industry and homeland security.