Event time:
Monday, April 28, 2014 - 12:00pm to 1:00pm
Location:
AKW 200
Speaker:
Jerrod Ankenman
Speaker affiliation:
Yale University
Event description:
Among the many important data analysis problems is the problem of matrices of high-dimensional data which has significant and possible non-linear correlations among different dimensions of the data. Examples include a collection of documents and the words in the documents, or a psychological questionnaire containing many questions and the responses of many people.
This talk will outline a data analysis algorithm for such matrices involving an iterative scheme for organizing each dimension by incorporating the organization of the other dimension, with a special focus on binary data which is modeled as independent Bernoulli samples from an underlying probability field.
Special note:
Non-standard day