#29166But now there are several commercial systems for the same language pair.
lr,5-5-P97-1015,ak
systems
</term>
. The method is based on
<term>
lists of words
</term>
from different
<term>
frequency classes
#29204The method is based on lists of words from different frequency classes.
other,10-5-P97-1015,ak
lists of words
</term>
from different
<term>
frequency classes
</term>
. It is shown how these
<term>
word
#29209The method is based on lists of words from different frequency classes.
other,16-7-P97-1015,ak
systems
</term>
that translate between
<term>
English
</term>
and
<term>
German
</term>
. It is challenging
#29243We also present the results of using our method on 6 MT systems that translate between English and German.
tech,11-7-P97-1015,ak
results of using our method on 6
<term>
MT systems
</term>
that translate between
<term>
English
#29238We also present the results of using our method on 6 MT systems that translate between English and German.
other,18-7-P97-1015,ak
translate between
<term>
English
</term>
and
<term>
German
</term>
. It is challenging to translate
#29245We also present the results of using our method on 6 MT systems that translate between English and German .
tech,13-6-P97-1015,ak
</term>
can be compiled and used for
<term>
testing
</term>
. We also present the results of
#29225It is shown how these word lists can be compiled and used for testing .
tech,6-1-P97-1015,ak
past the
<term>
evaluation
</term>
of
<term>
machine translation systems
</term>
has focused on
<term>
single system
#29144In the past the evaluation of machine translation systems has focused on single system evaluations because there were only few systems available.
tech,5-3-P97-1015,ak
</term>
. This requires new methods of
<term>
comparative evaluation
</term>
. In the paper we propose a
<term>
#29179This requires new methods of comparative evaluation.
tech,14-4-P97-1015,ak
the
<term>
lexical coverage
</term>
of
<term>
MT systems
</term>
. The method is based on
<term>
lists
#29196In the paper we propose a black-box method for comparing the lexical coverage of MT systems.
tech,20-1-P97-1015,ak
</term>
because there were only few
<term>
systems
</term>
available . But now there are several
#29158In the past the evaluation of machine translation systems has focused on single system evaluations because there were only few systems available.
tech,6-4-P97-1015,ak
</term>
. In the paper we propose a
<term>
black-box method
</term>
for comparing the
<term>
lexical coverage
#29188In the paper we propose a black-box method for comparing the lexical coverage of MT systems.
measure(ment),11-4-P97-1015,ak
black-box method
</term>
for comparing the
<term>
lexical coverage
</term>
of
<term>
MT systems
</term>
. The method
#29193In the paper we propose a black-box method for comparing the lexical coverage of MT systems.
tech,12-1-P97-1015,ak
translation systems
</term>
has focused on
<term>
single system evaluations
</term>
because there were only few
<term>
#29150In the past the evaluation of machine translation systems has focused on single system evaluations because there were only few systems available.
tech,4-1-P97-1015,ak
existing methods . In the past the
<term>
evaluation
</term>
of
<term>
machine translation systems
#29142In the past the evaluation of machine translation systems has focused on single system evaluations because there were only few systems available.
lr,5-6-P97-1015,ak
classes
</term>
. It is shown how these
<term>
word lists
</term>
can be compiled and used for
<term>
#29217It is shown how these word lists can be compiled and used for testing.
other,10-2-P97-1015,ak
commercial systems
</term>
for the same
<term>
language pair
</term>
. This requires new methods of
<term>
#29171But now there are several commercial systems for the same language pair.