PhDPrelimOralExam.pptVIP

09/20/2004 Parallelizing Exact SAR Model Solution Biography Education M.S., Electrical and Computer Engineering UMN-TC, 2000 Took WPE, Started PhD Thesis Ph.D. Candidate, Electrical and Computer Engineering UMN-TC expected 2005 Research Interests Data and knowledge engineering, spatial database management, spatial data mining, parallel processing, geographic information systems, and spatial statistics. Biography Publications High Performance Spatial Data-Mining, B. M. Kazar, S. Shekhar and D. J. Lilja, AHPCRC Tech Report no. 2003-125 Poster at ACM SIGPLAN Principles and Practice of Parallel Programming PPoPP Conference, June 2003. A Parallel Formulation of the Spatial Auto-Regression Model for Mining Large Geo-Spatial Databases, B.M. Kazar, S.Shekhar, D.J. Lilja, and D. Boley, AHPCRC Tech Report no. 2004-103, March, 2004. International Workshop on High Performance and Distributed Mining HPDM at SIAM Data Mining Conference, April, 2004. Comparing Exact and Approximate Spatial Auto-Regression Model Solutions for Spatial Data Analysis, B.M. Kazar, S. Shekhar, D. J. Lilja, K. Pace, AHPCRC Tech Report no. 2004-126 GIScience 2004 Conference, October, 2004 Scalable Parallel Approximate Formulations of Multi-Dimensional Spatial Auto-Regression Models for Spatial Data Mining, S. Shekhar, B. M. Kazar, D. J. Lilja, will appear as summary paper in 24th Army Science Conference Outline Spatial Data-Mining SDM Motivation Spatial Autocorrelation SA SDM Provides Better Model But, Computation Costs are Much Higher Computational Challenge Problem Definition Our Approach Algebraic Analysis Experimental Results Summary Future Work Motivation Widespread use of spatial databases Mining spatial patterns The 1855 Asiatic Cholera on London [Griffith] Fair Landing [NYT, R. Nader] Correlation of bank locations with loan activity in poor neighborhoods Retail Outlets [NYT, Walmart, McDonald etc.] Determining locations of stores by relating neighborhood maps with customer