iDE – um framework para suporte ao gerenciamento de cursos em educação a distância

Abstract

Managing distance learning courses (DE), offered in different locations and of difficult to access, is a mission that requires knowledge of the region, adaptation to local customs and a flexible pedagogical model that adapts to the characteristics of the classes. In this situation, the regional specificities and peculiar needs of the students form an exclusive knowledge of local teachers and tutors. And as such, these specialists use tacit knowledge to solve problems that arise during the course and that are specific to each location. These experiences become confined and restricted to the informal history of the class or course in focus. The reutilization of these solutions and their application in other similar situations would offer substantial benefits to the management of distance learning, as it would provide greater agility in solving issues with equivalent characteristics. This thesis presents the development of an intelligent framework, named iDE (intelligence for Distance Education), aimed at reutilizing solutions applied to the problems routinely faced by these courses in comparable scenarios. It combines techniques of Artificial Intelligence, such as Case Based Reasoning (CBR) and Multiagent Systems, along with information from an LMS and expert opinions, to retrieve and suggest solutions based on previous and similar circumstances. The iDE consists of a conceptual framework, managed by a multi-agent system, composed of four software agents: (i) the main agent and coordinator of the others (Manager agent); (ii) the agent that will use CBR techniques to recover and reuse knowledge of past cases (CBR agent); (iii) the agent that will process academic data from applications (mobile or not) inserted in virtual learning environments (LMS agent); and, finally, (iv) the agent responsible for recovering and formalizing the tacit knowledge of experts in the area of DE (Experts agent). A domain ontology was also modeled to represent specific data, and the instantiation of the Distance Learning Center of the Federal University of Amazonas (CED / UFAM), for exploratory case study purposes. The learning capacity of the framework was made possible by learning rules and by the direct evaluation by experts of the relevance of the returned answers. A web interface was also implemented to input data referring to the problem description and to the visualization of the responses suggested by the iDE. The actual information gathering process consisted of interviews with experts and the data collection of LMS Moodle. The validation experiments of the framework were performed with users with experience in DE, who evaluated the usefulness and relevance of the presented solutions.