SBIR-STTR Award

Data Mining Tool for Clustering Correlated Database Records
Award last edited on: 11/25/2002

Sponsored Program
SBIR
Awarding Agency
NSF
Total Award Amount
$100,000
Award Phase
1
Solicitation Topic Code
-----

Principal Investigator
Rohan Baxter

Company Information

Ultimode Systems LLC

2560 Bancroft Way Unit 213
Berkeley, CA 94704
   (510) 872-5238
   N/A
   N/A
Location: Single
Congr. District: 13
County: Alameda

Phase I

Contract Number: ----------
Start Date: ----    Completed: ----
Phase I year
1998
Phase I Amount
$100,000
This Small Business Innovation Research Phase I project from Ultimode Systems takes an innovative and mathematically sophisticated approach to clustering of large data sets in which records are linked through spatial or temporal vicinity, thus promising an original contribution to the advancement of clustering methodology. Clustering is a key method of data mining (knowledge discovery). Building on the success of ACPro, Ultimode Systems'data mining software tool which provides automatic clustering/segmentation for finding structure in high-dimensional databases, the project aims to invent effective representations and search algorithms to find clusters in databases containing sequentially and/or spatially related records, and to devise visualization tools for these more complex models to facilitate ease of application. Visualization of ACPro clusters provides insights into relationships and distinctions that would otherwise be difficult to identify in data. Many phenomena are most accurately represented as comprising database records correlated in time and/or space, but current clustering tools, including ACPro, treat database records separately and independently. Data mining or knowledge discovery, the discovery of patterns in large data sets, is increasing in importance in all branches of science and scholarly inquiry and in business. Large-scale imaging, instrument data collection, and the increase in electronic commercial transactions all create large data sets with spatial or temporal linkages. Thus improved methods for discovering patterns in such data sets will have significant impact in many areas by providing accurate modeling to a wider range of applications than has heretofore been possible, and, particularly through applicability in business, have the potential for commercial success. Existing clients working with correlated data, e.g., customer sales data, mineral exploration, and flight analysis, will benefit from adopting the resulting technology. In addition, the enhanced clustering capabilities will allow expansion into new markets where clients' databases contain correlated database records.

Phase II

Contract Number: ----------
Start Date: ----    Completed: ----
Phase II year
----
Phase II Amount
----