Cataloguing of learning objects using social tagging
Download as PPT, PDF0 likes346 views
The document discusses the cataloguing of learning objects using social tagging, exploring the concepts of visualization, recommendation, learning objects, and tag generation. It provides insights into the dynamics of social tagging and folksonomy, the importance of visualization techniques, and the application of recommendation algorithms in e-learning environments. The conclusion emphasizes the potential for developing applications that integrate these technologies to enhance the cataloging and recommendation of learning resources.
Cataloguing of learning objects using social tagging
- 1. Cataloguing of learning objects using social tagging XL Conferencia Latinoamericana en Informática Anderson Roque do Amaral, Luciana A. M. Zaina and José F. Rodrigues Jr. Universidade Federal de São Carlos – Sorocaba Available in: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6965111
- 2. Agenda introduction Motivation and Objectives Concepts Social web Folksonomy Learning Objects General Findings of Sistematic Review Social Tagging x Visualization Perspective Social Tagging x Recommendation Perspective Social Tagging x Learning Objects Pespective Social Tagging x Generation Tags Perspective Conclusions
- 3. Introduction Web Social Recent years have seen a proliferation of applications related Social Web, which are characterized by relating people and knowledge developed by the users. Social Tagging Social tagging in the Web allows the user to have freedom to create vocabularies that classify a given object. Folksonomy = folks(peoples) + taxonomy (classification) A folksonomy is the process of social tagging where users freely chosen keywords (also called tags) to identify and describe resources. Learning Environments Learning environments electronics has sought to adapt to the characteristics of this generation of Web applications to become more attractive and acceding to students
- 4. Motivation Search for new ways of cataloging, indexing and searching web content Challenge of applying new techniques emerged in the Web 2.0 in learning environments The dynamics of social tagging is based on the triad of People x Resources x Tags
- 5. Objectives Make a survey of existing solutions involving social tagging and learning objects Analyze the works found from 4 perspectives that we consider fundamental to social tagging in e-learning: Visualization, Recommendation, Learning Objects and Generation of Tags Identify potential topics of folksonomy to be explored in education area
- 7. Selected Works Articles in bold refer to Systematic Reviews A – Vizualization B – Recomendation C – Learning Objects D – Generation of Tags
- 8. Visualization Techniques In a social tagging or folksonomy system, the display of tags is crucial to support research, navigation and discovery of information for users. Studies have been developed to improve techniques that allow a visualization of the vocabulary and the relationships between tags.
- 9. Tag Cloud The tags are presented as a set of word sizes, different colors or fonts according to the weight of each, usually measured by frequency of use or popularity of the tags Elastic Maps Visualizes the emerging relationships between tags. The algorithm places the tags that are frequently used together in the same 2D plane. During the rollover effect, tags that tend to co-occur with the selected tag, are brought forward. Clicking on a tag, you can check the semantic context of the same
- 10. 6 PLi It is a visualization tool designed to be used with the site del.ico.us. Users can navigate through your own tags in an interactive network that employs different methods of 2D, 3D and circles image. You can choose the type of relationship between tags and resources are listed in the right part of the interface.
- 11. Other visualization techniques Concept Map; Hyperbolic visualization; Hierarchical diagrams; Multiple Inheritance; Evolution of tags.
- 12. Recommendation Perspective The technique of collaborative filtering is a technique that has been adopted in most recommendation systems of tags. It is based on filtering information based not only on the content of information but also in the assessment of people. Besides this technique, other algorithms have been applied on the recommendation of tags.
- 13. Recommendation algorithms Algorithm Authors Techinique PLSA Hofmann (1999), Cohn and Hofmann (2000), Jin(2004), Gui-Rong(2008), Arenas- García(2007), Hotho(2006) Wetzker(2009) Based on probability. Extension with Tags Tso-Sutter KHL (2008) adopts a model of non-directed graphs where nodes are tags and the edges are the relationships between tags FolkRank Algorithm Brin and Page(1998),Kleinberg(1999),Xi(2004),Hotho(20 06) adopts a model of non-directed graphs where nodes are tags and the edges are the relationships between tags Tensor factorization Symeonidis(2008),Rendle(2009) tag-based profile construction with a vector of weighted tags Noll e Meinel(2007),Diederich and Iofciu(2006), Stoyanovich(2008), Yeung(2008), Szomszor et al. (2007) Infer the user's interests through the tags most often used by him Naive Michlmayr(2007),Szomszor(2007),Dorigo and Caro (1999) is based on the frequency of use of tags WebDCC Godoy e Amandi(2006),Michalski and Stepp (1983),Thompson e Langley (1991). Recommendation by grouping tags Quadratic concept Jelassi et al(2012) QC = (U,T,R,P), Profile clusters
- 14. Learning Objects A learning object can be defined as an entity to be used within the teaching learning process. Among other things, we mention videos, pictures and simulators. Within the scope of e-learning what you want is to create digital content that can be reusable in different learning objectives The term Learning Objects (LO) was first used around 1992 by Wayne Hodgins, who through an analogy with pieces of 'lego', noted that individual objects represent knowledge that can be reused, shared and combined with other objects and each new combination new knowledge is created
- 15. Generation of tags In order to describe, understand and analyze tags and marking systems, several models were proposed generation of tags. These models study various factors that influence the generation of a tag, such as previous markings suggested by others, the knowledge of users, content resources and community influence. What sets these techniques of recommender systems is that objects that have not been marked, receive generated tags or invented by the system from some criteria.
- 16. Models Polya Urn. This model uses a simulation technique to capture previously assigned tags that are more likely to be selected again. The basic idea of the simulation is to place similar tags together in the same storage location (urn). In each simulation step, the selected tags are rearranged until a stabilization of vocabulary occur. The Simon-Yule model. the model is an extension of Polya Urn and its characteristic 'invent' new tags to add them into a stream of low probability. Thus, at each step of the simulation, it is verified which of the existing tags on the 'Urn' is more likely to come to be assigned to an object that was never marked by her
- 17. Conclusion It was observed through the study in this article that the prospects for visualization and recommendation generation of tags are little explored jointly. Combine the various technologies related to social tagging this work with the aim of favoring the cataloging, indexing, browsing and recommendation of the various types of learning resources in order to facilitate the whole process of teaching and learning is certainly an area that can be more explored. Applications that support the cataloging of learning objects through the social tagging can be developed by combining technologies that implement the four aspects presented in this review. These applications will support for experimentation where it is possible to analyze the process of cataloging learning objects in various sectors of education.
- 18. ACKNOWLEDGMENT The authors thank Centro Paula Souza , the State Government of Sao Paulo and CAPES for the financial support to this study