A92-1023 |
air travel information system (
|
ATIS
|
) . Air travel was chosen as
|
A00-2033 |
military simulation system . The
|
ATIS
|
grammar was extracted from an
|
A92-1023 |
either kind of answer . For the
|
ATIS
|
evaluations , two extensions
|
A92-1023 |
evaluations . As a result of the
|
ATIS
|
evaluations , a body of resources
|
A92-1023 |
Interpretation Document for the
|
ATIS
|
Application ( Note : these are
|
A92-1023 |
Specification ( CAS ) for the
|
ATIS
|
Application ( Note : this is
|
A92-1023 |
example , in the February 1991
|
ATIS
|
evaluations , the performance
|
A00-2033 |
non-left-recursive forms of the CT and
|
ATIS
|
gram - mars , but the size of
|
A00-1028 |
corpus of 15,000 trees from the
|
ATIS
|
domain using a version of the
|
A94-1010 |
models but not others for the
|
ATIS
|
domain . These results are consistent
|
A92-1023 |
would be counted correct . For the
|
ATIS
|
evaluations , it was necessary
|
A92-1023 |
to express the answers for the
|
ATIS
|
domain is presented in Appendix
|
A92-1023 |
Airline Travel Information Systen (
|
ATIS
|
) . This is a database of flights
|
A92-1023 |
re-implementec and modified by NIST for the
|
ATIS
|
evaluations . table for HYP ,
|
A92-1023 |
CAS specification for the DARPA
|
ATIS
|
evaluations , as distributed
|
A00-2033 |
non-left-recursive form of the
|
ATIS
|
grammar under the cut-off size
|
A00-2033 |
left-recursive productions . With the
|
ATIS
|
grammar and PT grammar , which
|
A92-1023 |
should be collected . For the
|
ATIS
|
evaluations , several different
|
A92-1023 |
them have been implemented in the
|
ATIS
|
evaluations ( Boisen et al. (
|
A92-1023 |
better scores from the February 91
|
ATIS
|
evaluation ) , and n = 145 ,
|