|
The Titan Project: Requirements for Data Mining &
Analysis.
Dunlavy
Wednesday, November 7th
1:15 - 2:15 pm
We discuss our current work in extending the Titan Informatics Toolkit to
handle text data. This work includes the development of a parallel data
ingestion and term-document matrix processing server. In the initial
development of the parallel text server, we will be integrating Epetra and
Anasazi for parallel matrix classes and SVD routines, respectively. We will
discuss some of the requirements of our project with respect to code
interface, data partitioning, data analysis, matrix decompositions, load
balancing, incremental data updates, parallel execution, and
more.
|