Transcript Slide 1 - Department of CSE

Educational Data Mining for Secondary and
Higher Secondary Education in Bangladesh
Md. Shafiqul Islam
d
Educational data is one of the huge resources of big data in
current world. This data record for our country is a great
source to extract information about our educational process,
progress and also an essential source of elements to predict
about the future learning behavior. But these utilities cannot
be gained merely looking over the raw data of education. A
proper procedure of analysis is the prerequisite to get
valuable information from these raw data, which is known as
Educational Data Mining (EDM).
Educational Data Mining refers to the techniques, tools, and
researches, designed for automatically extracting meaning
from large repositories of data generated by or related to
people's learning activities in educational settings.
Objective
Methodology
The first phase of the EDM process is to discover
relationships among data. This involves searching through a
repository of data from an educational environment with the
goal of finding consistent relationships between variables.
Our designed data warehouse repository is shown below.
Since both the repositories for SSC and HSC are almost
similar, here is only the schema for SSC is shown for
simplicity.
Dimension Table
Time
Fact Table
Fact_all
Session_start (pk)
Student_key (pk)
Roll
passing_year
Dimension Table
Location
Institution_code (pk)
Session_start (fk)
Subject_key (fk)
Reg
birth_day
Institution_code(fk)
birth_month
grade_point
birth_year
sex
group
institution
thana
Sylhet
Khulna
Chittagong
Rajshahi
Dinajpur
Barishal
Comilla
Dhaka
Humanity
Science
Business
Studies 2014
2010
2013 2011 2009
2012
2008
2005 2004
2006
2007
2003
2002
2001
Figure-2: A 3-D data cube representation of SSC/HSC data
according to Time, Location and Group. The result is shown with a
time domain of [2001-2014]. Though Madrasha and Technical board
does not shown, the corresponding result could be gain accordingly.
Expected Findings
Although educational data both in SSC and HSC are
extensive, these are much precise than many other
source of big data to examine for mining. Using the above
schema, we can perform the following tasks.
 Analyzing and visualization of student data
 Grouping students according to specific properties
 Detecting undesirable student behaviors
 Predicting student performance
The distilled data can also be used to plot into curve, bar
chart or into statistical regression analysis for human
judgment.
Student
Student_key (fk)
session_end
A typical implementation of our data warehouse schema is
illustrated in the figure below. A three-dimensional data cube
(3D Cube), is generated from data warehouse, which should
look graphically like the following diagram. The data here is
grouped into individual cubes according to Location, Group
and Time. Any relevant data cube can be generated from the
data available in the data warehouse, similarly. These cubes
should be observed to get information of educational records.
Time (Year)
Our objective is to design and implement an Educational
Data Warehouse repository, which may further be used to
extract useful information for Knowledge Discovery from
Data for educational data records (KDD). We have only
focused over the two public examination: one is Secondary
School Certificate (SSC) and the other is Higher Secondary
School Certificate (HSC) examination, in Bangladesh.
Dimension Table
Implementations
Group
Introduction
Dimension Table
area/district
Subject
sub_area/division
Subject_key (pk)
centre
subject_codes
subject_numbers
Figure-1 : Star Schema of SSC/HSC Data Warehouse.
Conclusion
In this thesis, we have focused on designing a data
warehouse to make the knowledge discovery from
educational data available for secondary and higher
secondary school examination data sources more easy.
Since the data is extensive and increasing with time,
challenges are associated with implementing educational
data mining. As a developing country Bangladesh also
needs to pay attention over the hidden information exists
into educational data resources for educational progress
and success.
References
1. Data Mining Concepts and Techniques(Third Edition) - by Jiawei
Han, Micheline Kamber, Jian Pei
2. http://www.educationaldatamining.org
3. http://en.wikipedia.org/wiki/Educational_data_mining
4. http://http://www.educationboard.gov.bd
Department of Computer Science and Engineering (CSE), BUET