recommender systems in e-learning environments€¦ · areas of presentations an overview of...
TRANSCRIPT
Recommender Systems in e-learning environments:
a survey of the state-of-the-art and possible extensions
PRATYA NUANKAEW
CORVI N US UN I VERS I T Y OF BUDA PEST, N OVEM BER 2 , 2016 .
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS1
Areas of PresentationsAn overview of Recommender systems in e-learning environments
◦ Introduction and Meaning of the recommender systems
◦ Challenges for designing a recommender system in e-learning environments
◦ A survey of the state-of-the-art in recommendation techniques for recommender systems
◦ A model for tagging activities and tag-based recommender systems
Some of the experiences with Recommender Systems◦ Determining of compatible different attributes for online mentoring model
◦ Online mentoring model by using compatible different attributes
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS2
Recommender Systems in e-learning environments: a survey of the state-of-the-art and possible extensions
Source: http://link.springer.com/article/10.1007/s10462-015-9440-z
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS3
Introduction and Meaning of the recommender systems
Recommender Systems (RSs) are software tools and techniques providing suggestions for items to be of use to a user (Kantor et al. 2011). The suggestions relate to various decision making processes such as what items to buy, what music to listen, what online news to read or what learning objects to learn.
Source: http://www.springer.com/us/book/9781489976369Citation: Kantor PB, Ricci F, Rokach L, Shapira B (2011) Recommender systems handbook. Springer, Berlin
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS4
Introduction and Meaning of the recommender systems
Personalization: learners have differences in background, goals, capabilities and personalities
Online learning activities: based on their preferences, knowledge and the browsing history of other learners with similar characteristics
e-Learning environments: ◦ each learner have (his/her) own tools, methods, paths, collaborations and processes.
◦ learning goal, prior knowledge, learner characteristics, learner grouping, rated learning activities (LAs), learning paths, and learning strategies, desired in a RS.
Results of the survey:◦ Advancement in the development of context-aware, e-learning recommenders
◦ e-Learning system from local learners to global learners
Source: http://www.springer.com/us/book/9781489976369Citation: Kantor PB, Ricci F, Rokach L, Shapira B (2011) Recommender systems handbook. Springer, Berlin
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS5
Challenges for Designing a recommender system in e-learning environments
Personalized recommendation approaches are first proposed in e-commerce area for product purchase (Balabanovi´ c and Shoham1 1997; Resnick and Varian2 1997), which help consumers find products they would like to buy by creating a list of recommended products for each given consumer (Cheung et al.3 2003; Schafer et al.4 2001).
The first challenge for designing a RS is to define the users and purpose of specific context or domain in a proper way (McNee et al.5 2006).
◦ The HRI Analytic Process Model. (Human-Recommender Interaction)
Citation:1Balabanovi´ c M, Shoham Y (1997) Fab: content-based, collaborative recommendation. Commun ACM 40:66-72
2Resnick P, Varian H (1997) Recommender systems. Commun ACM 40:56–583Cheung KW, Kwok JT, Law MH, Tsui KC (2003) Mining customer product ratings for personalized marketing.Decis Support Syst 35:231–243
4Schafer B, Konstan A, Riedl J (2001) E-commerce recommendation applications. Data Min Knowl Discov 5:115–1535McNee SM, Riedl J, Konstan JA (2006) Making recommendations better: an analytic model for human recommender interaction. In:
Conference on human factors in computing systems, Montréal, Québec, Canada, pp 1103–1108
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS6
The HRI Analytic Process Model. (Human-Recommender Interaction)5
Human-Recommender Interaction is a descriptive model, a way to understand recommenders from a different point of view. When added to a larger process model however, it becomes constructive—a way to analyze and redesign recommenders to better meet user information needs.
Citation:5McNee SM, Riedl J, Konstan JA (2006) Making recommendations better: an analytic model for human recommender interaction. In: Conference on human factors in computing systems, Montréal, Québec, Canada, pp 1103–1108
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS7
Challenges for Designing a recommender system in e-learning environments In a virtual classroom, teachers provide resources such as text, multimedia and simulations, and moderate and animate discussions.
◦ Web Usage Mining for a Better Web-Based Learning Environment (Zaïane 2001)1.
◦ Implement web learning environment based on data mining (Guo and Zhang 2009)2.
Source: http://roshnipatelsporfolio.weebly.com/web-based-learning-tools.htmlCitation:1Zaïane OR (2001) Web usage mining for a better web-based learning environment. In: Proceedings of conference on advanced technology for education, pp 450–4552Guo Q, Zhang M (2009) Implement web learning environment based on data mining. Knowl-Based Syst 22(6):439–442
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS8
Challenges for Designing a recommender system in e-learning environments
In e-learning environments utilizes, information about learners, learning activities (LA) and recommend items
◦ (Drachsler et al. 2008)1. presented “Personal recommender systems for learners in lifelong learning networks: the requirements, techniques and model.”
◦ based on previous learners’ activities or on the learning styles of the learners that are discovered from their navigation patterns.
To design an effective, it is important to understand specific learners’ characteristics (García et al. 2009; Chen et al. 2014)2:
◦ learning goal, prior knowledge, learner characteristics, learner grouping, rated learning activities (LAs),
◦ learning paths, and learning strategies, desired in a RSCitation:
1Drachsler H, Hummel HGK, Koper R (2008) Personal recommender systems for learners in lifelong learning networks: the requirements, techniques and model. Int J Learn Technol 3(4):404–4232García E, Romero C, Ventura S, de Castro C (2009) An architecture for making recommendations to courseware authors using association rule mining and collaborative filtering. User Model User
Adapt Interact 19(1– 2):99–1323Chen JM, Chen MC, Sun YS (2014) A tag based learning approach to knowledge acquisition for constructing prior knowledge and enhancing student-reading comprehension. Comput Educ
70:256–268
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS9
Challenges for Designing a recommender system in e-learning environments
E-learning systems should be able to recognize and exploit the learners’ characteristics as guidelines for implementation a good recommender system
• A good RS should be highly personalized
• A good RS should recommend materials at the appropriate time and location
• A good RS should be socially situated
• A good RS should include the adoption phase
• A good RS should support the continuous learning process
• A good RS should provide high level of interactivity
• A good RS should provide appropriate course materials according to learners’ (learning style)
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS10
A Survey of the State-of-the-Art in recommendation techniques for recommender system
E-learning system uses different recommendation techniques in order to suggest online learning activities to learners, based on their preferences, knowledge and the browsing history of other learners with similar characteristics.
Each recommendation strategy has its own strengths and weaknesses.
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS 11
Source: http://www.enkivillage.com/list-of-strengths-and-weakness.html
Recommendation Techniques for RS in e-learning environments
Each recommendation strategy has its own strengths and weaknesses.
In order to following material, it shows in the rest of the section.
This hierarchical structure includes all representative research examples.
◦ Matrix and tensor factorization methods
◦ Collaborative filtering approach
◦ Content-based techniques
◦ Association rule mining
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS 12
Source: http://ieeexplore.ieee.org/abstract/document/6934434/?reload=true
A Survey of the State-of-the-Art in recommendation techniques for recommender system
Matrix and tensor factorization methods
Focused on constructing recommender systems for recommending learning objects or learningactivities
Educational data mining has taken into account to support universities, teachers, and learners
The standard matrix and tensor factorization techniques◦ Predicting student performance (PSP)
◦ Mapping educational data to recommender systems
◦ Matrix factorization: implicitly encoding the “slip” and “guess” factors
◦ Tensor factorization for exploring the temporal effect
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS13
A Survey of the State-of-the-Art in recommendation techniques for recommender system
Collaborative filtering approach
Based on the assumption that learners has similar past behaviors and similar interests, then a collaborative filtering system recommends learning objects and the given learner have liked.
The examples of collaborative filtering system ◦ Altered Vista (AV) system
◦ Web-based PeerGrader (PG)
◦ The Scaffolded Writing and Rewriting in the Discipline (SWoRD) system
◦ Rule-Applying Collaborative Filtering (RACOFI) Composer system
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS14
A Survey of the State-of-the-Art in recommendation techniques for recommender system
Content-based techniques◦ Items are described by a common set of attributes.
◦ Learner’s preferences are predicted by considering the association between the item ratings and the corresponding item attributes.
Content-based techniques can be classified into two different categories1. Case based reasoning (CBR) techniques and,2. Attribute-based techniques
◦ Attribute-based Ant Colony System (AACS): There are three prerequisites for achieving
(a) The adaptive learning portal knows the learner’s attributes which include the learner’s knowledge level and learning style.
(b) The learner’s attributes and learning object’s attributes which have been annotated by teacher or content providers.
(c) Matching the relationships between learners and learning object
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS15
A Survey of the State-of-the-Art in recommendation techniques for recommender system
Association rule mining◦ Ways of representing discovered knowledge and describe a close correlation between frequent items in a
database
◦ In e-learning systems, aims to intelligently recommend on-line learning activities to learners based on the actions of previous learners
Examples:◦ Agents for on-line learning activities • Automatically leading the learner’s activities
◦ Identifying attributes of performance • Discovering interesting learner’s usage information
◦ Finding out the relation among the learning materials • Finding learners’ mistakes
◦ optimizing the content of an e-learning portal • Deriving useful patterns
◦ Personalizing e-learning based on comprehensive
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS16
A Model for Tagging Activities and Tag-Based Recommender SystemsCollaborative tagging is employed as an approach, which is used for automatic analysis of user preference and recommendation.
A model for tagging activities
Social tagging systems allow their users to share their tags of particular resources.
◦ Figure shows a conceptual model for social tagging system where users and items are connected through the tags they assign.
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS 17
A Model for Tagging Activities and Tag-Based Recommender SystemsTag-based recommender systems
◦ Recommender systems in general recommend interesting or personalized information objects to users based on explicit or implicit ratings.
◦ Personalized recommendation use to conquer the information overload problem
◦ Tagging represents an action of reflection
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS 18
Source: http://www.slideshare.net/KarenLi27/tag-based-recommender-system
A Model for Tagging Activities and Tag-Based Recommender Systems
Applying tag-based recommender systems to e-learning environments◦ The innovation with respect to the e-learning system
◦ Learners could benefit in two important ways: ◦ Tagging is proven a meta-cognitive strategy
◦ Learners’ tags could give comprehensible recommendations about the resources
The following features of collaborative tagging to success in e-learning1. The information provided by tags
2. Enhance peer interactions and peer awareness centered on learning content
3. Reflective practice
4. A lack of the social cues
5. Tagging provides possible solutions for learners’ engagement
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS19
Conclusions• Contributes to the conceptual and theoretical understanding of RS in e-learning environments
• Highlights the important requirements and challenges for designing a recommender system in e-learning environments
• Limited research about collaborative tagging in education, growing interests in exploring and unlocking the value of the increasing meta-data
• Tagging provides possible solutions for learners’ engagement
• Collaborative tags, created by learners to categorize learning contents, would allow instructors to reflect at different levels on their learners’ progress
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS20
Some of the experiences withRecommender Systems
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS21
Determining of Compatible Different Attributes for Online Mentoring Model
Source: http://ieeexplore.ieee.org/abstract/document/6934434/?reload=true
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS22
Conceptual Diagram of Online MonitoringThe mentee requires the appropriate mentor
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART AND POSSIBLE EXTENSIONS 23
Attributes Definition
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS24
Attributes Definition
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS25
Relationships of mentor and mentee attributes
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART AND POSSIBLE EXTENSIONS 26
Satisfaction Value Defining
Online Mentoring
Model
Satisfaction Defining
CorrespondentsSurvey
StatementHypothesis
Compatible Different Attributes
Questionnaires for Mentee
Mentee
[588 students]
Satisfaction Value toward Mentee
attributes
5 statements of Hypotheses for
mentee
Mentee’s point of view
One statements is the same for
mentor and mentee.
Different point of view
Questionnaires for Mentor
5 statements of Hypotheses for
mentor
Mentor’s point of view
Mentor
[74 lectures]
Satisfaction Value toward Mentor
attributes
The
mat
chin
g m
od
el s
ho
uld
be
dev
elo
ped
fo
r o
nlin
e
mo
nit
ori
ng.
Attributes Definition
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS28
The satisfaction rating
The satisfaction values are analyzed by using Net Promoter Score (NPS).
The respondents are divided into three groups including: I. Promoters: (9 – 10) that agree within a high level
II. Passives: (7 – 8) that general satisfaction, but not positive which they actively.
III. Detractors: (0 – 6) that dissatisfaction tend to disagree with the proposed topic
Net Promoter Score (NPS) Calculation
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS 30
The resultsThe proposed attributes is tested with 74 lectures and 588 students from Faculty of Information and Technology, Rajabhat Maha Sarakham University to reveal their satisfaction on the proposed attributes.
They are asked to complete the survey questionnaire to obtain the satisfaction value.
We found that the same attribute might not obtain the same satisfaction values from different points of view (Mentor and Mentee).
This paper proposed the new attributes for online mentoring.
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS31
Online Mentoring Model by Using Compatible Different Attributes
Source: http://link.springer.com/article/10.1007/s11277-015-2755-x
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS32
Attributes Analysis
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS33
Attributes Analysis
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS34
Signification and Attitude toward attributes
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS35
Components of compatible different attributes
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS36
Satisfaction Tool Developing Process
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS37
Virtual community
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS38
Thank you.
Nov 1 2016RECOMMENDER SYSTEMS IN E-LEARNING ENVIRONMENTS: A SURVEY OF THE STATE-OF-THE-ART
AND POSSIBLE EXTENSIONS39