International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-8, Issue-1, May 2019
2414
Published By:
Blue Eyes Intelligence Engineering
& Sciences Publication
Retrieval Number: A1961058119/19©BEIESP
Abstract: Educational Data Mining (EDM) explains the
exploration involved with the application of data mining,
machine learning and statistical analysis to the enormous
amount of data generated from educational institutions. At a
high level, the sector seeks to develop and improve strategies for
exploring this information, which frequently has multiple levels
of significant hierarchy, so as to find new insights regarding the
learning process of individuals in the context of such settings.
Therefore, the EDM has contributed to theories of learning
investigated by researchers in educational psychological science
and the learning methodology. This sector is closely tied with the
learning analytics, which are compared and contrasted. This
work is a comparative analysis of various decision tree
classification algorithms using Engineering students’ academic
performance data. Educational Data Mining is the process
which extracts knowledge through interesting patterns
recognized from large amount of data from educational field.
Learning related datasets with the performance of students
obtained from educational institutions and processed before
actual data mining or data analytics process. Data mining is one
of the information discovering regions which is broadly used in
the field of computer science. Furthermore is an
inter-disciplinary area which has great impact on various other
fields such as data analytics in prediction of risk factors in
business organizations, medical forecasting and diagnosis,
market basket analysis, statistical analysis and forecasting,
predictive analysis in various other fields. Data mining has
multiple forms such as text data mining, web data mining, visual
data mining, spatial data mining and Educational data mining.
As educational institutions is the source of generating quality
students in order to tune them to become an eminent personality.
All the educational institutions must be aware of the competency
and academic level of every student in order to upgrade their
performance. The implementation work is performed in Weka
tool to compare the performance accuracy between the different
types of decision tree classification algorithms namely J48,
Entree and Enhanced Random Tree. These three classifier
algorithms which are widely working with the Weka tool are
used to classify this learning dataset and the result are obtained
and has been evaluated & compared to identify the best decision
tree classifier among them.
Index Terms: Educational Data mining; Weka, decision tree,
classifier, Learning dataset, J48, RepTree, Enhanced Random
Tree.
Revised Manuscript Received on May 31, 2019
Dr. A.S.ARUNACHALAM, Department of Computer Science, School of
Computing Science, VISTAS, Chennai, Tamil Nadu, India.
Mr. A. THIRUMURTHI RAJA, Department of Computer Science,
School of Computing Science, VISTAS, Chennai, Tamil Nadu, India.
Dr. S.PERUMAL, , Department of Computer Science, School of
Computing Science, VISTAS, Chennai, Tamil Nadu, India.
I. INTRODUCTION
Educational Data Mining is an imminent trend in the
higher education. The quality of the student’s utilization and
enhancement of variety of learning techniques used in
Educational Institutions may be improved through
discovering hidden knowledge from the large amount of data
stored in student databases/ data warehouses. In general
the process of data mining has many tasks from
pre-processing. The actual task of data mining starts after the
pre-processing task obtained raw data in to a processed one
so as to apply the data mining techniques [1]. Sturdy patterns
if found can doubtless generalize to create correct
predictions on future knowledge. The amount of data in the
medical field has increased tremendously. Although, such a
large volume of information is valuable and need to be
analyzed for further forecast perceive and predict the
complexities that may arise in future. Data mining gives the
methodology and technology to recognize the valuable
information of data for higher cognitive process. Machine
learning algorithms are widely used to analyses and process
various kinds of data [2].There are various tool available to
apply the machine learning algorithms to different kinds of
data set. These tools allow performing various kinds of tasks
from pre-processing till visualization of the obtained result
[3]. In this work the Weka tool is used for implementing and
evaluating the machine learning algorithms using the
medical dataset.
II. REVIEW OF LITERATURE
Educational data mining is a major area mainly focus for
predicting the students’ academic performance in l earning
aspects. The extraction of necessary information from
collected educational record set and analyzing the
information are known as educational data mining. The
evaluation of research field and recent improvement in
educational filed leads to produce a challenging task in
evaluating students’ performance in academics. The
necessary steps for educational data mining starts from
pre-processing following feature extraction process and ends
with analysing stage with necessary clustering and
classification algorithms [23]. Data mining techniques are
broadly divided into: supervised and unsupervised learning
[24,29]. Data mining techniques are pertained to forecast
students’ academic performance based on some of the
attributes like socioeconomic condition, earlier academic
performances and so on.
Enhanced Constructive Decision Tree
Classification Model for Engineering Students
Data
A.S. Arunachalam, A. Thirumurthi Raja, S.Perumal