Novel Software Tools for Analyzing Massive Datasets

The overall objective of this proposal is a feasibility analysis of developing a software package for using topologically based ideas to process and extract information from large data sets. The project will consist of four main components: (1) a comprehensive review of the literature concerning topological methods related to the analysis of large data sets (2) identification of additional promising topological and geometric theory that may result in practical algorithms (3) evaluation of algorithms and prospects for numerical implementation including parallel versus sequential programming considerations (4) a detailed design of the software package that includes an evaluation of the suitability of various platforms including Matlab, R and C++.