Information Technology Reference
In-Depth Information
5. ProPoSEd frAmEWork
in India, Hyderabad) will show error once they
are evaluated by machine as well as by a English
spoken man. Ex 1- Prime Minister Manmohan
Sing Garu has visited Osmania University. Ex 2
- Hyderabad says albida to monsoon. Where the
'Garu' is a pure Telugu word and used in English
newspapers published form Andhra Pradesh.
'Albida' is an Urdu word and very much used in
English newspapers coming out form Luckhnow
and Hyderabad. Local languages influence same
as English used in Chiana, Japan, Korea or France
and no one considers them as Error. Where as from
English point of view they are wrong. Of course
a good number of such words got chance to be
included in Oxford dictionary. Research shows
present AEG systems illustrate 10 - 15% lower
score while using Indian English text as input.
In a broader form it can be mentioned that the
English spoken and written by non-native English
people (i.e. - Asians) are very much influenced by
local languages. India is a multilingual country
with as many as 22 scheduled languages and only
5% (plz. read five percent!!) of the population is
able to understand English and that's also not like
USA or UK people. Hence a goal is to develop a
framework for an AEG system (for those countries
where English is not a common language), which
can be used for correcting essays influenced by
local languages, and also to teach how to write
better English Essays. In this chapter I propose a
standard framework to develop any Automated
Essay Grading System under such context. This
model can be executed or s/w can be build as per
the requirement as for example in which part of
world the system is going to be used. Because
while writing English the students of Japan are
not influence by Hindi or Tamil, it is Japanese.
Hence a single system will not be able to solve
the problem. But this framework can be used as
a benchmark to develop the other AEG systems.
The framework follows IEEE Std. 1471 -2000,
which is about “IEEE Recommended Practice for
Architectural Description of Software Intensive
Systems”.
Under the above circumstances a need of a spe-
cialized AEG system was felt very much which
can be used for engineering education and we can
make it web based also. The system will have the
capability of identifying the local languages (Ex -
Indian) present in the submitted essay and it will
also find out how much effect is there for these
words. It will also help the students to resubmit
the essay with corrections where the students will
be asked to re-enter the similar words for those
local languages. Their essays will be graded as it
is they have entered the equivalent English words
by their own. For the instructors or teachers it will
also give a proper score-card mentioning that still
how much the essay is influenced by the local lan-
guages and the no of local words present, number
of times corrections made by students (i.e. they can
be given two or three chances to enter equivalent
English words for those local words (i.e. albida =
good bye). This above-mentioned action is a part
of the scoring engine. These functionalities are
added as a new functional module in the scoring
engine or score reporter.
The feedback module is also supported with a
'local language' engine which helps the students
providing proper feedback and development
notes along with English Grammatical mistakes
improvement, fed back on use of too many weak or
common words etc. This engine will be very much
useful in the learning stage. At the very begin-
ning this engine will identify the local languages
present in the written essay. Then it will give a
chance to the students to overcome this problem
by providing equivalent English words by their
own. Then it will show them the projected score
with number of general (English) errors and pres-
ence of number of local languages and what are
those. For the remaining local words in the essay
the system will then suggest equivalent English
words with similar English words of those. Now
the students get a chance to substitute remaining lo-
cal words, phrases with suggested English words.
Search WWH ::
Custom Search