REALIZATION OF MACHINE ALGORITHM FOR THE DETECTION IN TEXT OF OBJECT NAMES,
FOREIGN TECHNOLOGY DIV WRIGHT-PATTERSON AFB OHIO
Pagination or Media Count:
The problem of detecting and separating names of objects in text consists of finding in text by phrase testing names of objects, described by the given text, and in the obtainment of a list of these names. Objects in text may be indicated a by natural names b by changes in positions v predicates acquiring the name value in given context. The problem is solved in two stages 1 separation of object names a and b by comparing each word of the processed assumption with words from a special listing, stored in OZU ZTSVM, 2 with the aid of a recognizing assumption form and special criterion separating names of objects of b type. The program for solving this problem was realized on an URAL-4 machine. Objects can be indicated in text by the following types of names a natural names, b local names, c predicates, acquiring name values in given context. If the list has the name of type v, and further for a corresponding object was detected a new name of type a, then the name of type v is crossed out from the list and in text to word form, corresponding to the name of type a, is put down a marker.