Higher Education in Pakistan: An ICT Integration ViewpointZaffar Ahmed Shaikh
More Related Content
Similar to SkillsRec: A Novel Semantic Analysis Driven Learner Skills Mining and Filtering Approach for Personal Learning Environments based on Teacher Guidance
Similar to SkillsRec: A Novel Semantic Analysis Driven Learner Skills Mining and Filtering Approach for Personal Learning Environments based on Teacher Guidance (20)
If this Giant Must Walk: A Manifesto for a New Nigeria
SkillsRec: A Novel Semantic Analysis Driven Learner Skills Mining and Filtering Approach for Personal Learning Environments based on Teacher Guidance
1. SkillsRec: A Novel Semantic Analysis Driven Learner Skills
Mining and Filtering Approach for Personal Learning
Environments based on Teacher Guidance
Authors: Zaffar Ahmed Shaikh, Denis Gillet, Shakeel Ahmed Khoja
Presenter: Zaffar Ahmed Shaikh
2. Agenda
• The Problem
• Our Solution
• Abstract
• Introduction
• Related Work
• Results
• Conclusions
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 2Friday, March 27, 2015
3. Teacher guidance
Problem Solution
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 3Friday, March 27, 2015
4. Abstract
• SkillsRec is a novel teacher guidance based learner skills mining and
filtering approach that identifies learner skills for PLE based learning
scenarios using Latent Semantic Analysis (LSA) technique.
• SkillsRec is developed on PLE design and development principles of
the guided PLEs model [1].
• This paper compares learner-skill similarity scores generated through
the SkillsRec with those generated through conventional IR and KM
techniques.
• We provide top N=8 user-user recommendations most likely to be
similar for a given active learner.
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 4Friday, March 27, 2015
5. Introduction / PLE
• Online PLE is a modern day personalized learning based
teaching/learning environment.
• PLE can be defined as “a highly flexible ‘one-size-fits-all’ solution to
online learning that provides personalized, collaborative, inquiry-based
and guided learning experiences to Internet users [4]”.
• PLE takes care of learner personality, mood, interests, and needs during
her interaction with online environments [3].
• In addition, PLE concept addresses information overload problem
through recommender technology [5,6].
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 5Friday, March 27, 2015
6. Introduction / gPLEs model
• The gPLEs model incorporates teacher-based guidance mechanism into
the PLE concept through learner skills mining and filtering based
recommendation mechanism [1].
• It identifies/develops learner skills through semantically analyzing teacher
competencies [2] and learner interests.
• Using those skills of a learner, it provides her with skill-similarity based
user-user recommendations.
• The model has been / can be implemented in online PLE(s).
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 6Friday, March 27, 2015
7. Introduction / LSA
• LSA is a model-based natural language processing and data/document
retrieval technique that improves retrieval process through developing
measures of semantic similarities between user and text [7].
• LSA performs various statistical computations to search items that
match with user query based on contextual usage meaning of words in
user query and item descriptions [7,8].
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 7Friday, March 27, 2015
8. Introduction / SkillsRec
• SkillsRec is a model-based learner skills identification and assessment technique for CF based
recommendation systems. It works on descriptive/unstructured data.
Mines user data to
identify user skills
Filters user skills matching
with teacher roles
Generates user-user
recommendations in ranked
order
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 8Friday, March 27, 2015
9. Related Work
• Model-based techniques (Bayesian models or LSA) have been used
before in modeling user profile and her context [10,11,12].
• In existing literature there is no evidence about finding learner skills
through analyzing learner interests against teacher roles.
• There is also a lack of information in literature about exploiting learner
interests-related data for generating similarity recommendations.
• Hence, the main mission of this work was to develop a CF based
recommendations system which employs natural language processing
tool (LSA) to identify learner skills which are based on teacher guidance.
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 9Friday, March 27, 2015
10. Data Organization
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 10Friday, March 27, 2015
11. Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 11Friday, March 27, 2015
12. Data Analysis
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 12Friday, March 27, 2015
13. Results
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 13Friday, March 27, 2015
14. Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 14Friday, March 27, 2015
15. Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 15Friday, March 27, 2015
16. Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 16Friday, March 27, 2015
17. Conclusions
• We have presented here the SkillsRec–a novel semantic analysis based
recommender model for PLEs.
• This provides the solution to how to overcome the massive, exponentially
increasing [9], information overload problem.
• SkillsRec provided results have been compared with conventional IR and KM
based similarity techniques.
• It can be concluded from the presented details and results that semantic
analysis based data mining and filtering approaches provide promising
results; thus, they need to be further explored and tested.
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 17Friday, March 27, 2015
18. References
[1] Shaikh, Z.A., Khoja, S.A.: Towards Guided Personal Learning Environments: Concept, Theory, and Practice. In:
14th IEEE International Conference on Advanced Learning Technologies, pp. 782-784. IEEE Press, New York
(2014).
[2] Shaikh, Z.A., Khoja, S.A.: Personal Learning Environments and University Teacher Roles Explored using Delphi.
Australasian J. Educ. Tech. 30, 202-226 (2014).
[3] Moore, M.G.: Transforming e-learning. Keynote Address to The 3rd International Conference on e-Learning
and Distance Learning, Riyadh, Saudi Arabia (2013).
[4] Gillet, D.: Personal Learning Environments as Enablers for Connectivist MOOCs. In: IEEE International
Conference on Information Technology Based Higher Education and Training, pp. 15. IEEE Press, New York
(2013).
[5] El Helou, S., Gillet, D., Salzmann, C.: The 3A Ranking System: Contextual, Personalized & Simultaneous
Recommendation of Actors, Activities & Assets. J. Universal Comput. Sci., Special Issue on Context-Aware
Recommendations, (2010).
[6] Bogdanov, E., Ullrich, C., Isaksson, E., Palmér, M., Gillet, D.: Towards PLEs through Widget Spaces in Moodle.
Comput. Sci. Inf. Syst. 11, 443-460 (2014).
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 18Friday, March 27, 2015
19. [7] Landauer, T.K., McNamara, D.S., Dennis, S., Kintsch, W.: Handbook of Latent Semantic Analysis,
Psychology Press, (2013).
[8] Kalz, M, van Bruggen, J., Giesbers, B., Waterink, W., Eshuis, J., & Koper, R..: A study about placement
support using semantic similarity. Educational Technology & Society, 17 (3), p. 54-64 (2014).
[9] Hofmann, T.: Collaborative Filtering via Gaussian Probabilistic Latent Semantic analysis. In: 26th ACM
SIGIR, pp. 259-266, (2003).
[10] Verbert, K., Duval, E., Lindstaedt, S.N., Gillet, D.: Context aware Recommender Systems. J. Uni. Comp.
Sci. 16, 2175-2178, (2010).
[11] Verbert, K., Manouselis, N., Ochoa, X., Wolpers, M., Drachsler, H., Bosnic, I., Duval, E.: Context-aware
Recommender Systems for Learning: a Survey and Future Challenges. IEEE Trans. Learn. Tech. 5, 318-335
(2012).
[12] Drachsler, H., Verbert, K., Duval, E.: Recommender Systems for Learning. 1-20, Springer. New York
(2013).
[13] Chien, Y.-H., George, E.I.: A Bayesian Model for Collaborative Filtering. In: Online Proceedings of The
Seventh International Workshop on Artificial Intelligence and Statistics, (1999).
[14] Baloian, N., Galdames, P., Collazos, C., Guerrero, L.: A Model for a Collaborative Recommender
System for Multimedia Learning Material. In: Vreede de, G.-J., Guerrero, L. Marn Ravents, G. (eds.) CRIWG
2004. LNCS, vol. 3198, pp. 281-288. Springer Berlin Heidelberg (2004).
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 19Friday, March 27, 2015
20. [15] Herlocker, J.L., Konstan, J.A., Terveen, L.G., Riedl. J.T.: Evaluating Collaborative Filtering
Recommender Systems. ACM Trans. Inf. Syst. 22, 5-53, (2004).
[16] Lonsdale, P., Baber, C., Sharples, M., Byrne, W., Brundell, P., Beale. R.: Context Awareness for
MOBIlearn: Creating an Engaging Learning Experience in an Art Museum. In: MLEARN, pp. 115-118
(2004).
[17] Cui, Y., Bull, S.: Context and Learner Modelling for the Mobile Foreign Language Learner. System. 33,
353–367, (2005).
[18] Graesser, A.C., Wiemer-Hastings, P., Wiemer-Hastings, K., Harter, D.: Using Latent Semantic Analysis
to Evaluate the Contributions of Students in Autotutor. Inter. Learn. Environ. 8, 129-147 (2000).
[19] Zampa, V., Lemaire, B.: Latent Semantic Analysis for User Modeling. J. Intell Info. Sys. 18, 5-14
(2002).
[20] Wolfe, M.B.W., Goldman, S.R.: Use of Latent Semantic Analysis for Predicting Psychological
Phenomena: Two Issues and Proposed Solutions. Behavior Research Methods Instruments & Computers.
35, 22-31 (2003).
[21] Laham, D., Bennett, W., Landauer, T.K.: An LSA-based Software Tool for Matching Jobs, People and
Instruction. Inter. Learn. Environ. 8, 171-185 (2000).
[22] Wu, L., Hoi, S.C., Yu, N.: Semantics-preserving Bag-of-words Models and Applications. IEEE Trans.
Image Process. 19, 19081920, (2010).
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 20Friday, March 27, 2015
21. Questions?
Zaffar Ahmed Shaikh – EPFL /IBA Karachi - zaffarahmedshaikh@gmail.com –AINA 2015 (MAW’15) – March 25 2015 21Friday, March 27, 2015
Editor's Notes
The conventional IR technique applies Cosine similarity on descriptive/unstructured data while traditional KM technique uses keywords/structured data. The aim is to report SkillsRec gains over conventional IR techniques.