Pattern Analysis and Modeling.
Abstract
The report summarizes research progress in new approaches to search and extraction of structure in patterns which include a nondirectional parallel search algorithm which has been shown to be better than the alpha-beta search algorithm when dealing with game trees. New methodology for the design of hierarchical statistical classifiers includes use of state-space ordered search algorithms and introduction of heruistics to cut down the number of decision rules to be tried in a dynamic programming approach to optimizing the decision policies at the nodes of a decision tree. Also described are contributions to the modeling of communication channels with memory and the investigation of models of database query languages. Among the papers published on the research performed under this grant, the following two papers present substantive overviews of the research and place the work in the perpective of other contributions in the areas discussed. Both papers include extensive lists of references. The two papers are: Laveen N. Kanal, 'Problem-Solving Models and Search Strategies for Pattern Recognition' IEEE Trans. on Pattern Analysis and Machine Intelligence, Volume 1, Number 2, p. 193-201, April 1979; and Laveen N. Kanal, and A. R. K. Sastry, 'Modeling Channels with Memory and their Applications to Error Control', Proc. of the IEEE, Volume 66, Number 7, p. 724-744, July 1978.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1979
- Accession Number
- ADA070961
Entities
People
- Laveen N. Kanal
Organizations
- University of Maryland