Automatic Informative Abstracting and Extracting
Annual progress rept. Mar 1969-Mar 1970
LOCKHEED MISSILES AND SPACE CO INC PALOALTO CA PALO ALTO RESEARCH LAB
Pagination or Media Count:
Part I of this report documents several experiments in automatic indexing. Nine chapters, each from a different technical book were used as the text copies for all the experiments. In the first experiment, an attempt was made to construct a sentence dictionary of syntactic sentence types, for distinguishing extract-worthy sentences, but it proved unrewarding. Algorithms developed to combine syntactic and statistical criteria in the choice of extract sentences and index phrases proved more rewarding. Extract sentences and index noun phrases from several texts are presented for the reader to peruse. There is every indication that satisfactory back-of-the-book indexes could be produced automatically, with post-editing to delete superfluous items. Part II reports on the relationship between English word government and the problem of multiple meaning in material language processing. A set of English words is discussed, each of which has the ability to distinguish among semantic meanings by the use of certain syntactic units such as prepositions. Because prepositions play an important role in making semantic distinctions, a section on prepositional semantics is included.
- Information Science
- Computer Programming and Software