Statistics Colloquium

Event Type: 
Emily Casleton
Event Date: 
Thursday, April 25, 2019 -
3:30pm to 4:45pm
SMLC 356

Event Description: 

Title: Multi-Source Data Fusion and Imputation Methods



Multi-source data provides a unique opportunity to learn about a single phenomenon from

different perspectives. Specifically, for the context of facility monitoring, various, disparate

sensors can monitor different aspects of a facility, and events of interest or patterns of life will

manifest differently in the various data streams. In this talk I will discuss an overview of multisource

data fusion in the context of a variety of projects I have worked on at Los Alamos. One

issue that is commonly encountered when analyzing multi-source data is missing data.

Although this issue is not exclusive to multi-source data analysis, it is more likely to occur with

multiple sensors and can lead to discarding a large amount of data collected. However, the

information from observed sensors can be leveraged to estimate those values not observed. I

will discuss two methods for imputation of multi-source data, both of which

take advantage of potential correlation between data from different sensors, through

ridge regression and a state-space model. Performance of imputation methods are compared

with the mean absolute deviation; however, rather than using this metric to solely rank the

methods, I will also present an approach to identify significant differences. Imputation

techniques will also be assessed by their ability to produce appropriate confidence

intervals, through coverage and length, around the imputed values.

Event Contact

Contact Name: Li Li

Contact Email: