A framework for construction of
learning process using
data mining based approach
Guided by Presented by
Dr. R Santhi V.Vanitha
ABSTRACT
Important components of an e-learning system is the learning object.
Reuse -High quality learning materials are expensive to create.
Learners benefiting from e-learning depends on the extent to which
appropriate learning objects are sequenced.
Difficulties in retrieving the suitable LO
Incomplete metadata and/or incorrect metadata
Also due to vast amount of LO, it is impractical to hand.
Data mining can be used as the potential tool.
Propose a framework that uses data mining algorithm
for the discovery of appropriate learning objects.
The objective of the project is to
Define a framework for the construction of
Learning process using Learning Objects.
Extend the framework to incorporate data mining
based search tool for the retrieval of Learning
Objects.
OBJECTIVE
INTRODUCTION
A Self-contained learning content
Ideally, each covers a single topic
Serve as building blocks for
lessons, modules, or courses
Can be reused in multiple
instructional contexts
LEARNING OBJECT
A Self-contained learning content
Ideally, each covers a single topic
Serve as building blocks for lessons,
modules, or courses
METADATA
Purpose -support the reusability
The IEEE 1484.12.1 2002 Standard
is an internationally-recognized open
standard.
Learning
Object
LO Metadata
INTRODUCTION
LEARNING PROCESS/LEARNING PATH
Sequence of the learning objects to build a well structured
and a well designed course.
Creation of Learning Path.
Usage of LO to create learning
Process is a key factor.
Previous LO indicate prerequisite topic.
Enhanced Learning Through Learning Knowledge Object
BY
Sabitha etal., IJETCAS, 2013
Objective :
Search and delivery of LO to learner
Approach:
Creation of LO and KO and metadata
Extraction of LO and KO for user query based on metadata
Convergence of LO and KO (LKO) - classification technique.
Delivery of LO by clustering LKO.
Drawbacks:
Retrieval of LO obj_id, topic, subtopic,date, author, type
Not suffice Educational Level
Mining implicit - depend on ability & expertise
Knowledge
Application of Data Mining Techniques to Regulated Learning System
BY
Eleonora Brtka etal., IEEE SISY, 2010
Objective :
Use of Data mining techniques to deliver appropriate LO to learner
Approach:
Combines IEEE LOM and IMS LIP attributes
Uses Rosetta tool (based on rough set theory) to generate
association rules.
Based on rules, Los are retrieved.
Drawbacks:
Missing/incomplete /inconsistent attribute Extraneous rule
IMS LIP standard is optional Not all LMS implement
Several attributes Numerous rules and plenty of
LOs
Applying Data Mining For Predicting Student Performance
BY
Edin Osmanbegovi etal., Journal of Economics and Business, 2012
Objective :
Effectiveness and advantages of Data mining techniques in the
..domain of education.
Research question:
What are main advantages of data mining and e-learning
interaction
Outcome:
Support the specific needs of participants in the educational
process
All participants will be benefited.
Application - prediction of students' academic
performances.
Challenges facing the retrieval and reuse of Learning Object
BY
Harris Michael etal., ECDL, 2010
Objective :
Investigate challenges facing the retrieval and reuse of learning
..Objects.
Discussion:
Organizational issues Due to competition, no sharing.
Cultural issues Insecurity, loss of appreciation.
Technical issues format for storage, incomplete metadata.
Conclusion
ability to find a Learning Object is also affected by the
way it has been indexed.
the better metadata, the better the chance
to be found.
Embedding and validating Empirical Usage Intelligence in Learning
Objects
Nugent etal., EDMEDIA, 2011
Objective :
Impact of LO on students learning
Discussion:
Study the perspective of 850 students and 27 teachers
Students interest (M = 4.25 on 5-point scale)
Better understand the topic (M = 3.97).
Good-Excellent (M = 4.31).
attractiveness, control over learning, ease of use, interactivity.
Conclusion
Learning objects can be successfully used for
independent learning of complex subject matter.
EXISTING SYSTEM
The learning objects are retrieved based on
Based on Learning Object Metadata Subjective, human error.
Learning profile or/and learner preference Recommender
Learning style of the learner Not applicable
Automatic extraction of certain metadata from document
using text mining approach - suitable to specific type
PROPOSED SYSTEM
LOM is created out by Instructional designers manually
Subjective - Inaccurate - Incomplete
Discoverability and reusability of LO depends on LOM
Developing a framework that
Creates a Wrapper around the metadata based on how LO is
being used.
Complete metadata All relevant attributes are considered.
Accurate metadata Automatically extracted.
Sequence the extracted Learning object to create Learning
process
.
SYSTEM ARCHITECTURE
DATA FLOW DIAGRAM CONTEXT LEVEL
CONTEXT LEVEL DFD DIGARAM
LEVEL 0 DFD DIAGRAM
DATA FLOW DIAGRAMLEVEL 1
(Design Module) .
DATA FLOW DIAGRAMLEVEL 2
(Search LO) .
MODULES
Design and implementation of Database Layer
Design of LO and LOM
Building LOR
Design and implementation of User Interface
Interface for students and tutors
Design and implementation of Application logic
Building integrated Process
Retrieval of LO using Association rule mining
MODULES
WORK DONE IN PHASE 1
Design and implementation of Database Layer
Data modeling techniques ERD
Identifying entities
Defining relationship between entities
Identifying cardinalities
Defining PK, attributes,FK
Mapping Object model and Relational data model
Design & implementation of Application Logic & respective User
Interface
Implementing prototype
Login, Upload LO, view LO, Search LO using Keyword,
build LP
Students view their course materials.
WORK DONE IN PHASE 1
WORK DONE IN PHASE 1
SCREEN SHOT OF DATABASE DIAGRAM
SCREEN SHOT OF LOGIN
SCREEN SHOT OF SEARCH LO
REFERENCES
Gregg.G., E-learning agents, The Learning Organization Vol. 14 No. 4, 2007,pp.
300-312 Emerald Group Publishing LimitedPrakasam.s., An agent -based
Intelligent System to enhance E-Learning through Mining Techniques,
International Journal on Computer Science and Engineering Vol. 02, No. 03,
2010, 759-763
Miller.LD, etal., iLOG: an Artificial Intelligence Framework for Automatic
Metadata Generation for Online Learning Objects, internatioal Journal of
Artificial Intelligence in Education 19 (2009) .
Nugent.G , Kupzyk.K, etal., Empirical Usage Metadata in Learning, 39th
ASEE/IEEE Frontiers in Education Conference, October 18 - 21, 2009, San
Antonio.
Pouyan.P, etal., An Agent Based Multilayered Architecture for E-learning System,
The Second International Conference on E-Learning and E-Teaching (ICELET
2010).Yaghmaie .M, Bahreininejad.A., A context-aware adaptive learning
system using agents, Expert Systems with Applications 38 (2011) 32803286.
THANKYOU !!!
DATA FLOW DIAGRAM CONTEXT LEVEL
DATA FLOW DIAGRAM LEVEL 0.
WORK DONE IN PHASE 1- AUTHORING LAYER
WORK DONE IN PHASE 1- COMPOSITION
WORK DONE IN PHASE 1- COMPOSITION
IEEE LOM
Problems with manual tagging
Friensen (2009)
- - discussed difficulty of incorpoarting incorrect meta
tagged LO into design
Cechinel (2010)
- - Study on human error in indexing the LO.
Roy (2008)
- - Less efficient, more costly and less consistent.
Meta data tagging and search
ABSTRACT
One of the most important components of an e-learning system
is the learning object or the learning content. High quality
learning materials are expensive to create. So it is very
important to ensure reuse of learning content. There are
practical difficulties in retrieving the suitable LO due to
incomplete metadata and/or incorrect metadata. Also due to
vast amount of LO, it is impractical to hand pick the appropriate
data. Learners benefiting from e-learning depends on the extent
to which appropriate learning objects are sequenced. Data
mining can be used as the potential tool to provide solution to
this problem. The purpose of this master thesis is to propose a
framework for the design of learning process through data
mining. The framework uses data mining algorithm for the
discovery of appropriate learning objects based on what learner
is required to know.
