L1 - Based Major Component Detection and Analysis (L1 MCDA) for n-Dimensional Data Clouds Based on Linear Conic Programming
Abstract
The proposed work will seek to establish new theory and efficient algorithms for identifying the major componcnt(s) of statistically distributed data clouds in n-dimensional Euclidean space. principally tluough use of the Ll norm. The proposed research will uncover new theory and solution/approximation methods of linear conic programming to develop L1 Major Component Detection and Analysis (MCDA) for higher dimensional data clouds that has "all operations in L1" and direct connection with heavy-tailed statistics. It will proceed by: - Efficient computation of central point as the initial iterant of the n-dimensional Ll MCDA, - Fast calculation of characteristics of radial extension in Step 2 of then-dimensional Ll MCDA, - Conduct computational experiments of the n-dimensional Ll MCDA for higher dimensional data clouds - Explore and develop new models and efficient computations based on linear conic optimization theory.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 12, 2017
- Source ID
- W911NF1510223
Entities
People
- Shu-cherng Fang
Organizations
- Army Contracting Command
- North Carolina State University
- United States Army