H. Kargupta, E. Johnson, E. Riva Sanseverino, H. Park, L. D. Silvestre, and D. Hershberger. Scalable data mining from distributed, heterogeneous data, using collective learning and gene expression based genetic algorithms. Technical Report EECS-98-001, School of Electrical Engineering and Computer Science, Washington State University, 1998.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....optimal solution grows exponentially with the problem size in absence of the knowledge regarding the important partitions. The genetic algorithm community have realized the importance of linkage learning in the scalability of the genetic algorithm. A growing number of linkage learning techniques [3, 4, 12, 23, 20, 24, 31, 30, 38, 39, 51, 52, 65] are becoming available. Most of these techniques are based on a combination of heuristic based and statistics based approaches. The history of linkage learning e orts dates back to Bagley s dissertation [2] Bagley used a exible representation and the so called inversion operator for adaptively ....

....This is a heuristic based approach and often, its application is computationally expensive. The fast messy GA [22] o ered some reduction in computational cost of linkage learning. Several new techniques have come up in the recent past. The gene expression messy GA (GEMGA) introduced elsewhere [4, 31, 30, 38] o ered a technique for linkage learning. The GEMGA makes use of a local perturbation technique to identify the genes that are critical for high tness. Once the genes are identi ed in the local context, they are put together in clusters and tested for global performance. Good clusters de ne the ....

H. Kargupta, E. Johnson, E. Riva Sanseverino, H. Park, L. D. Silvestre, and D. Hershberger. Scalable data mining from distributed, heterogeneous data, using collective learning and gene expression based genetic algorithms. Technical Report EECS-98-001, School of Electrical Engineering and Computer Science, Washington State University, 1998.

....currently under development, addresses all of the above issues. Although additional facets of DDM systems are likely to emerge in the future, we believe the characteristics, listed above, o er a good starting point for developing a DDM system. For example, the BODHI system was rst reported in [20], and since that point in time it has undergone signi cant design changes. Some of these changes came from practical issues that arose during the implementation process, and others out of further study and re ection upon the problems being addressed. 6.1 Design Principles The BODHI System is an ....

H. Kargupta, E. Johnson, E. Riva Sanseverino, H. Park, L. D. Silvestre, and D. Hershberger. Scalable data mining from distributed, heterogeneous data, using collective learning and gene expression based genetic algorithms. Technical Report EECS-98-001, School of Electrical Engineering and Computer Science, Washington State University, 1998.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC