© IJARW | ISSN (O) - 2582-1008
September 2021 | Vol. 3 Issue. 3
www.ijarw.com
IJARW1525 International Journal of All Research Writings 11
SECURE INTELLIGENT ESSAY SCORING SYSTEM (SIESS)
USING END-TO-END ENCRYPTION
Madubuezi Christian Okoronkwo
1
, Caroline Ngozi Asogwa
2
, Nwuorgu Raymond
3
1,2
Department of Computer Science, Michael Okpara University of Agricultural, Umudike
3
Computer Science Department, University of Nigeria, Nsukka
ABSTRACT
The study addressed the issues associated with the conventional grading which can be biased, prone
to human error and usage of inefficient scoring attributes as well as subjectivity in essays grading. The
research work has been able to demonstrate the effort of Natural Language processing (NLP) in
developed Secure Intelligent Essay Scoring System (SIESS). Natural language processing is concerned
with utilization of knowledge of Computer Science. The essay scoring apply several methods, a Holistic,
Analytic and Primary trait. The method used for this was combining the holistic and analytic method.
The study made use of an online platform called kaggle.com used for a pre-build model as a python
library. Also, the essay scoring system was evaluated using the confusion matrix majorly the accuracy
metric which is specifically recommended for machine learning algorithm (Naïve Bayes, Linear
Regression and Gradient Boosting). The Gradient Boosting was best approach essay scoring pre-built
model prediction. Datasets in excel template were equally collected in the domain such as operating
system, expert system, database management, data structure and algorithm. SQLite as the database
management system was used for data storage. The system was tested in a local hosting architecture
capable of running online too. The entire system is summarized in three concepts. Mechanism of
teachers, students and automatic registration.
Keyword: Cryptography, Encryption, Automated Scoring System, Natural Language Processing
1. INTRODUCTION
Machine Learning has become generally embraced
in an assortment of genuine applications which
altogether influence individuals' lives (Guimaraes
and Tofighi, 2018; Guegan and Hassani, 2018) in
Automated Essay Scoring (AES) particularly when
sent as a component of high stakes tests.
Instructive "applications dependent on Natural
Language Processing (NLP) and Automated Essay
Scoring (AES) Systems frequently utilize a blend
of Machine Learning calculations like straight
relapse (Manvi, Mishel, and Ashwin, 2012),
(Ramalingam, Pandian, Prateek, and Himanshu,
2018), counterfeit neural organization (Marwa,
Rania, and Ahmed, 2020), gathering procedure
(Arif, Byung-Won, Ingyu, and Gyu, 2017) in
consequently reviewing understudies' essay and
this have demonstrated to be a decent match or
override human grader unwavering quality. An
automated essay scoring (AES) program is a
product that utilizes procedures from corpus and
computational phonetics; and machine learning to
grade essays. In schooling, composing capability is
a fundamental part of correspondence and
deciding if, an understudy can pass on a message
or express a thought with coherence and riches.
Composing capacity appraisal has progressed
from understudies noting numerous decision
undertakings to built reaction errands. In which
understudies are needed to peruse a prompt(s),
ponder significant central issues, and foster an
essay; subsequently, looking at their language use,
composing abilities, and basic reasoning. As far as
evaluating composing capacity, there is an overall
need to select and prepare human raters to
allocate scores. However, this course of manual
scoring has its impediments like weariness,
interruption, irregularity of scoring across time,
etc (Leyi, Yali, and Yan, 2019). This wasteful and
thorough method of scoring has achieved the
requirement for quick reaction including moment
scoring and investigation of composing quality.
Thus the rise of Automated essay scoring (AES)
framework, to arrangement composing scores by
PC programs. Some Automated Essay Systems go
past their extent of offering all encompassing
scores, to likewise giving reaction respect to
sentence design, association and language use, etc.
The inescapable reception of automated scoring
systems and organization innovations, have