METRIC SPACE FOR FRAMES AND NEURAL NETWORK CLASSIFIERS
Abstract structures and generalisations are useful in order to solve common practical problems in A.I. Frames are structures representing stereotyped situations and each frame contains different types of information. A metric space structure is introduced in the set of frames. If A(r) is a reference frame containing information about a reference subject R, and Ac(i) (i=1,...,N) a number N of "candidate frames" C(i), representing subjects similar but not equal to R, a classical problem is the selection of the frame C(j) representing a subject, the most similar to subject R. A procedure which can do this will be a Neural Network Classifier.
Bibliographic Reference: PAPER PRESENTED: EURASIP WORKSHOP ON NEURAL NETWORKS, SESIMBRA (PT), FEBRUARY 15-17, 1990 AVAILABLE FROM COMMISSION OF THE EUROPEAN COMMUNITIES, DG XIII-C-3, L-2920 LUXEMBOURG AS PAPER EN 35200 ORA
Availability: Can be ordered online
Record Number: 1989128086900 / Last updated on: 1990-11-01
Original language: EN
Available languages: en