SOCIO-AFFECTIVE MODULE FOR RECOMMENDER OF COMPETENCY LEARNING OBJECTS MSA-RECOACOMP: a study in development Walber Lins Pontes Patricia Alejandra Behar Magda Bercht Federal University Rio Grande do Federal University Rio Grande do Federal University Rio Grande do Sul - UFRGS Sul - UFRGS Sul - UFRGS Caixa Postal 5071 - 90.041-970 Caixa Postal 5071 - 90.041-970 Caixa Postal 5071 - 90.041-970 Porto Alegre - RS - Brazil Porto Alegre - RS - Brazil Porto Alegre - RS - Brazil +55 98 3265-0103 +55 98 3265-0103 +55 51 3308-6817 walberpontes@gmail.com pbehar@terra.com bercht@inf.ufrgs.br ABSTRACT With the technological advance, new conceptions of teaching and This article describes the required parameters for the learning emerged as the embodiment of the OAs in face-to-face development of the socio-Affective Module (MSA) of education or distance learning. Recommender of learning objects by competencies Such features provide easy access to the themes under (RECoaComp)-MSA-RECoaComp. This is intended to recognize consideration, enabling the subject engaging independently and the socio -affective aspects in recommending Learning Objects autonomously. (OAs) skills. The module is being implemented by a multidisciplinary team and is on the prototyping phase. In the Given this, one of the challenges of the educator is the selection first stage were scaled the elements that will support the socio- and organization of these materials in order to contemplate the affective recognition process. Such data will be extracted by profile of its students and their needs. MSA-RECoaComp an exisiting environment of distance education The availability of content which is not suited to the needs and and is used at the institution, ROODA more specifically one of its socio-affective characteristics of students cause an overload of resources, the Affective Map [14], and the Recommender of information to the user. competency Learning objects (RECoaComp). Thus, this work allows you to understand the functionality of the MSA- As suggested solution of this problem is the recommendation RECOACOMP noting the feasibility of the recommendation systems, according to [5] are intended to assist the user in the regarding the OAs filtering skills considering the socio-affective search and selection of content focused on profile, working aspects. literally as information filters. Categories and Subject Descriptors Thus, the user receives as a result of searches only the closest and relevant materials, as long as using and feeding system with new K.3.1 [Computers and Education]: Computer Users in information, either to the profile or research it carries out. Education - Collaborative Learning. The recommendation system that this article is about objectively General Terms filters the OAs according to skills considering the socio-affective Performance, Human Factors, Verification. recognition of users. Note that this article covers aspects of characterization of the Keywords module developed and does not discuss issues relating to the Socio-affective recognition; Recommendation skills; Meaningful system itself, considering that it is still being modeled. learning. The recommendation system has the potential to collaborate on indication of OAs more suitable socio-affective aspects of the 1. INTRODUCTION user, being an important tool both for classroom education as the This article describes the structure of the socio-affective distance in different educational contexts. Such a structure is recognition on recommendation of learning objects based on based on moods and motivational factors of [21] and on the skills. sociogram [17]. This article is organized into six sections and section one introduction. Section two introduces the concept of Such a feature is being developed by a multidisciplinary team in OAs, the section three the OAs recommendation. Section four - stages, which is a socio-affective recognition. Five - features related systems that preliminary result of work identifying the socio-affective aspects will support the MSA. Section six presents the socio-affective to be considered for the MSA in RECoaComp filtering process. aspects to be implemented in the structure of the MSA and their perspectives for recommendation. 2. LEARNING OBJECTS This approach will give emphasis on collaborative filtering, The teacher uses the Learning Objects (OAs) to mediate content filtering and the hybrid, by supporting the process of information in knowledge construction, with Wayne Hodgins the recommendation skills present in the RECoaComp. first to use the term in 1994 [8] Collaborative filtering is based on systems that perform the [23] conceptualizes OA as possible digital resources to be reused process of recommendation through the human assistance, to support teaching. [22] broadens the understanding by resulting from the collaboration of groups interested in that acknowledging the OA as any additional feature to the learning element. It has limitation on the recognition of the interest and process, by unlinking it from the need to be a digital element. It understanding of the individual contributor on the object, as well has five characteristics: 1) the information that must be next to the as on the recommendation process itself. object (metadata); 2) reusability; 3) accessibility; 4) Content-based filtering is constituted as systems that apply the interoperability; and 5) durability, presented as rules to recognition of elements that can have common interest implicit or standardize the development of objectives so that they serve the explicit. The process may happens by distinct approaches, but characteristic of reuse. with main purpose of recommendation. Considering the need of reuse, granularity of OA and their Hybrid filtering recognizes the possibility of interacting more than availability in stores it is necessary to recognize its features and one filtering technique allowing the simultaneous use of two or functionalities. This way, the recommendation systems allow you more, in order to be provided the limitations of each mode [1] and to find something inside the large set of OA (s) that (s) he can be [5]. re-used effectively. 4. SOCIO-AFFECTIVE CAPTURE ON 3. RECOMMENDATION OF LEARNING RECOMMENDATION OF COMPETENCY OBJECTS The recommendation process considering multiple alternatives for LEARNING OBJECTS On the perspective of recommendation of OAs by skills, a solution. In the search for the most suitable choice generally considering the socio-affective recognition,three elements are performs a direct choice, or through previous recommendations considered for the student's interaction with the OA: 1) socio- [16]. affective space; 2) motivational factors and 3) State of mind. He thus considers the need to recommend content, elements or information matching the expectations of the individual [21]. 4.1 Socio-Affective Space [6] emphasizes the challenge of recommendation systems to The socio-affective space is being considered from the concepts of perform the appropriate combination between expectations of [18], when he establishes the sociogram structure based on social users and the elements to be recommended. interactions. The Sociogram, is a graphical representation of sociometry, and allows the identification of group interactions, or formation of social networks, the establishment of groups and the 3.1 Recommendation skills highlights or reference elements as well as the marginalized The skills-based recommendation takes into account the need to elements within the social structure. assist the user in the search and selection of focused content to the profile [6]. This process is not characterized as a filtering system, but a guideline for interaction of the recommendation system. 4.2 Motivational Factors The motivational factors are developed from Bercht model [4] In this context it is relevant to understand the great challenge of with influence from [9] who considers the independence, the the recommendation is to recognize the combination of elements effort and the student's confidence in execution of tasks and that make possible an appropriate result to the expectations of the activities in a virtual system. The three elements to be combined users. subsidize the inference of motivation, being a hint for the The choice of filtering process gains importance as it identifies the recognition of the State of mind. characteristics of the recommendation and the needs of This work was used entirely in [15] when considering evaluation individuals involved in the process. The modeling of the system of motivational factors a persistent set of actions adopted by the becomes critical to contemplate the most reliable results possible student in the Virtual learning environment (VLE). to offer or need incorporated. 3.2 Filtering systems 4.3 State of mind Within the context of recommendation seven types of filtration The mood is based on definitions of [22]: 1) be excited, implies a systems are described: 1) collaborative filtering; 2) content-based joyful behavior demonstrate good mood, motivation, interest, filtering; 3 demographic filtration); 4) knowledge-based filtering; satisfaction to meet the challenges of learning, and collaborates 5) utility-based filtering; 6) based filtering in other contexts; 7) and cooperates with partners; 2) be discouraged, implies hybrid filtering. demonstrate a discontent, sad behaviour, unwilling, disinterest, without motivation, dissatisfaction, frustration (or feel penalized) The first two systems are observed in the texts of [11] and [21]; to continue learning, or even feel coerced, by believing that the the third has highlighted in the text of [17]; the fourth and fifth will of others prevails; 3) be indifferent, implies demonstrate are found in [5]; the sixth is approached [12] and [19]; the apathy, carelessness, negligence, neglect and lack of motivation seventh is a result of the above found in [1] and [5]. for learning content. 5. SYSTEMS TO BE USED IN THE MSA- RECoaComp template Macro vision RECOACOMP Aspects of categorization for the socio-affective recommendation of competency learning objects will be recognized and made available by systems validated by the core of Digital Technology applied to education (NUTED): 1) Map, affective ROODA functionality, and the 2) RECoaComp [3]; and [7]. 5.1 ROODA The ROODA, institutionally recognized by UFRGS in 2003 as E- LEARNING environment. It is the AVA in this work as a platform chosen for the implementation of framework of recognition and validation of socio-affective States of the students during the OAs recommendation processes. The ROODA aims the main educational paradigm shift from the Figure 2. Cazella et ali, 2012. interaction and cooperation of users in AVA. User-centered and In general the basic operation of RECoaComp happens in three value-driven process of cooperation. For [2], the goal of this steps: 1) the teacher selects OAs from a repository, aiming at the platform is to offer possibilities through resources on the web. construction of specific skills, recognizing that it can supply more Users (teachers, counselors and students) can build a cooperative than one jurisdiction; 2) the student responds to a questionnaire work through virtual and social interactions, turning your way of which traces a profile about the competencies relevant to the thinking from the coexistence and exchange between students and subject (these previously defined by professor); 3) is triggered the teachers. search through the information filtering by selecting the default repository, using the registered metadata, the OA with the student 5.1.1 Affective Map profile, regarding competences [8]. The Affective Map [15] is a feature of ROODA which considers The idea of RECoaComp is to provide the student the content that the moods of students and was developed in four phases as shown best meet the needs of building skills based filtering at the in Figure 1, inspired by [13]: a) acquisition and identification; b) intersection of information relating to the student's profile and interpretation; c) selection and d) inference of the moods of the skills that make it possible to develop OA. student. The acquisition and identification determine the means and methods by which the system will recognize characteristics relating to affective States under review. 6. MSA-RECoaComp And Prospects The MSA will be developed in the form of a RECoaComp General scheme of recognition of moods module. Its implementation happens with the identification of the State of mind and motivational factors obtained through the affective and social relations maps presented on the sociogram. It is of growing interest the development of technological tools directed to educational systems that deal with the recognition of social and affective phenomena. It discusses, in this work, the introduction of affective aspects and sociometry, based on mood, motivational factors and sociogram, in virtual learning environments, to facilitate the provision of OAs by competencies. It is intended to apply the experiments during the semester of 2015/I in students of the Business Course and later, in students in the Post- graduation Program in Informatics in education. The application will check the student's perception about the appropriateness of the OAs when recommended within the socio- affective aspects. Figure 1. Longhi et ali (2007) The study is relevant to determine whether the variables chosen for the delineation of the categories of State of mind and social 5.2 RECoaComp and its perspectives environment should be considered (or reassessed) when developing tool MSA-RECOACOMP.. The RECoaComp [3] allows the filtering of OAs considering the competences to be established by the user. Using a collaborative 7. REFERENCES filtering and content-based filtering, characterizing it as a hybrid [1] ADOMAVICIUS, G.; TUZHILIN, A. (2005). Toward the Filtering that assists in indication of relevant materials to the Next Generation of Recommender Systems: ASurvey of the State- student's profile. The RECoaComp was developed according to of-the-Art and Possible Extensions. IEEE Transactions on the model of Figure 2. Knowledgeand Data Engineering, New York, v. 17, n. 6p. 734- [13] LIAO, W.; ZHANG, W.; ZHAU, Z.; JI, Q.; GRAY, W. D. 749. Toward a decision-theoretic framework for effect recognition and [2] ALVES, D. O teste sociométrico. Rio de Janeiro, FGV, 1964. user assistance. International Journal of Human-computer Studies, v. 64, n9, p. 847-873. 2006. [3] Behar, Patricia Alejandra ; Silvio Cazella ; KELLEN, Ketia ; SCHNEIDER, Daisy ; FREITAS, R. . Developing a Learning [14] Longhi, M. T., BERCHT, M. BEHAR, P. A. Objects Recommender System based on Competences to Reconhecimento de Estados Afetivos do Aluno em Ambientes Education: experience report. In: The 2014 World Conference Virtuais de Aprendizagem. Revista Novas Tecnologias na on Information Systems and Technologies (WorldCIST 14), Educação. Porto Alegre,V. 5, N. 2 Dez. 2007. 2014, Ilha da Madeira. New Perspectives in Information Systems [15] Longhi, M. T. Mapeamento de aspectos afetivos em um and Technologies. Heidelberg: Springer, 2014. v. 1. p. 217-226. ambiente virtual de aprendizagem. 2011. 253f. Tese [4] BERCHT, M. Em direção a agentes pedagógicos com (Doutorado: PGIE), Centro Interdisciplinar em Novas dimensões afetivas. 2001. 152f. Tese (Doutorado: Programa de Tecnologias na Educação, UFRGS, Porto Alegre, 2011. Pós Graduação em Computação), Instituto de Informática, [16] MAES, P.; SHARDANAND, U. (1995). “Social information Universidade Federal do rio Grande do Sul - UFRGS, Porto filtering: Algorithms for automating "word of mouth”, In: Human alegre, 2001. Factors in Computing Systems. Proceedings…, 1995, p. 210- [5] BURKE, R. (2002). Hybrid recommender systems: Survey 217. and experiments. User Modeling and User-Adapted Interaction, [17] MONTANER, M., LÓPEZ, B., DE LA ROSA, J.L. (2003). Dordrecht, v. 12, p. 4, p. 331–370. A Taxonomy of Recommender Agents on the Internet. Artificial [6] CAZELLA, S. C.; REATEGUI, E.; MACHADO, M.; Intelligence Review. Netherlands : Kluwer Academic BARBOSA, J. (2009) Recomendação de OAs Empregando Publishers, pp. 285-330, Aug. Filtragem Colaborativa e Competências. In: Simpósio Brasileiro [18] MORENO, Jacob L. Quem sobreviverá?. Goiânia: de Informática na Educação (SBIE). Dimensão, 1994. [7] CAZELLA, S. C.; SILVA, K. K. da; BEHAR, P. A.; [19] NUNES, M. A. S. N. (2009). Recommender Systems based SCHNEIDER, D.; FREITAS, R. RENOTE, Recomendando on Personality Traits:Could human psychological aspects Objetos de Aprendizagem baseado em Competências em EAD. influence the computer decision-making process?. 1. ed. Berlin: Porto Alegre, v.9, n.2, 2011. VDM Verlag Dr. Müller. v.1. 140 p. [8] CAZELLA, S. C., BHEAR, Patricia, SCHNEIDER, Daisy [20] Nunes, M. A. ; Cazella, S. C. (2011), “O que sua Schneider, SILVA, Ketia Kellen da, FREITAS, Rodrigo. personalidade revela? Fidelizando clientes web através de Desenvolvendo um Sistema de Recomendação de OAs baseado Sistemas de Recomendação e traços de personalidade”. In: em Competências para a Educação: relato de experiências. Patrícia Vilain; Valter Roesler. (Org.). Tópicos em banco de Anais do 23º Simpósio Brasileiro de Informática na Educação dados, multimídia e Web. Tópicos em banco de dados, (SBIE 2012) Rio de Janeiro, 26-30 de Novembro de 2012 multimídia e Web. 1ed. Florianópolis: Sociedade Brasileira de [9] CEdMA: Learning Architectures and Learning Objects. Computação, v. 1, p. 91-122. Disponível em:. Acesso [21] RESNICK, P. e VARIANT, H. R. (1997) Recomender em: abr. 2014. systems. Comunications of the ACM, New York, V. 40, n. 3, pp. [10] Del SOLDATO, T.; Du BOULAY, B. Implementation of 55-58, mar. motivational tactics in tutoring systems. Journal of Artificial [22] SCHERER, K. R. What are emotions? And how can they be Intelligence in Education, v. 6, n. 4, p. 337-338. 1995. measured? Social Science Information, v. 44, n4, p. 675-729. [11] GOLDBERG, D., NICHOLS, D., OKI, B. M., TERRY, D. 2005. (1992). Using collaborative filtering to weave an information [23] TAROUCO, Liane; FABRE, Marie-Christine Julie Tapestry. Communications of the ACM, New York, v.35, n.12, Mascarenhas; TAMUSIUNAS, Fabrício Raupp. Reusabilidade de p. 61-70, Dec. objetos educacionais. Revista Novas Tecnologias na Educação. [112] GONZALEZ, G., DE LA ROSA, J.L., and MONTANER, Porto Alegre, p. 1-11. 2003. M. (2007). Embedding Emotional Context in Recommender [24] WILEY, D. A. Learning objects and the new CAI: So what Systems. In The 20th International Florida Artificial do I do with a learning object?. 1999. Disponível em: Intelligence Research Society Conference-FLAIRS, Key West, . Acesso em: 02 de março de 2014.