Page 110 - ASBIRES-2017_Preceedings

Alagalla & Jayakody
as less completion rate of courses and
difficulty to follow courses schedules by
students. There are no such tools available
to analyze the MOOC data at the beginning
stage of the courses to identify the root
causes of the above mentioned problem.
Therefore, this research focused on
developing an efficient solution by using a
dataset which consists of 600000 records
from the Harvard/MIT MOOC’s system.
The main focus of this research is
to analyze the root causes of students’
failures by data analysis of MOOC’s
courses through the Graph analysis and
Data analysis. Graph analysis was
implemented to analyze more than two
data attributes at once accurately by
Hadoop big data framework and its
packages. Moreover, Data analysis was
implemented to get the accuracy of the
MOOC’s results by Python and its machine
learning packages.
As a result output, this system gave
the most affected attribute and its
probability of influence to the student
results. In addition, this system can be used
to improve the MOOC’s system to be a
productive tool for student centric courses
and improve student’s results.
2 LITERATURE REVIEW
2.1 Impacts of MOOCs on Education
Impact of MOOCs on education
system is discussed more broadly.
Specifically it focuses on the disruptive
potential of MOOCs, and so a large portion
of the scholar’s review is relevant to this
issue. For instance, she identifies
characteristics of some MOOCs such as
their size, automation in grading, and their
openness (particularly with regard to
cMOOCs) as factors with the potential to
affect approaches to teaching and learning.
The size and openness of MOOCs are also
highlighted by Kennedy (2014) in terms of
their potential to “[disrupt] conventional
thinking about the role, value, and cost of
higher education” Gasevic, Kovanovic,
100
Joksimovic and Siemens (2014) examined
numerous problems associated with
MOOC. The data collection and data
analysis methods of MOOC research are
areas that are poorly investigated in prior
literature.
3 METHODOLOGY
The research data set was mainly
taken from Harvard and MIT MOOC’s
system. It consists of 600000 records in the
dataset. The research methodology was
broken into two sections such as graph
analysis and data analysis.
Figure 1: Architectural diagram of system
3.1 Graph Analysis
Graph analysis was designed
through the Hadoop framework. Dataset
consists of numerous attributes to analyze
such as number of days activated users,
number of videos seen by users, number of
discussions gone through by forums,
number of chapters covered, number of
events gone through by users etc. Hadoop
framework used to analyze data by graphs.
Backend of the methodology was
developed with hive, hBase and pig.
(Phaneendra & Reddy, 2013, Mridul,
Khajuria, Dutta, Kumar, 2014, Gasevic,
Kovanovic, Joksimovic & Siemens, 2014).
Map reduce process was used by the
Hadoop framework to generate the
graphs.Map reduce process
1. Prepare the Map() input – the "Map
Reduce system" designates Map