Analítica de aprendizaje y personalización
Loading...
Publication date
Authors
Advisors
Department
Research group
Center
Abstract
En este trabajo se plantea una revisión de las analíticas masivas de datos de aprendizaje en la Educación Superior. Se relaciona con un nuevo paradigma de aprendizaje basado en tareas y en logros en consonancia con las capacidades individuales y no con el tiempo, con el espacio o con la edad. La viabilidad y la relevancia las define claramente el problema de 2 sigma, que plantea el amplio horizonte por recorrer hasta un objetivo límite de aprendizaje. Actualmente se constata un decidido interés por el análisis de datos de aprendizaje utilizando los sistemas y el software basado en los entornos sociales y ubicuos y en los nuevos LMS que lo incorporan. El problema es que hasta ahora las herramientas consolidadas de uso común solo obtienen datos y gráficas que relacionan el rendimiento individual con el grupal, y el de éste en conjunto, y además sólo lo hacen con referencia a datos de aprendizaje que hemos introducido merced a procedimientos de evaluación convencionales. Sin embargo hay un espacio que suministra una enorme cantidad de datos no solo para la evaluación del alumno y que actualmente ignoramos, al menos de forma explícita, es el espacio de trabajo personal del alumno conectado, en red con sus iguales, con los profesores, con los recursos y con todo el material que va utilizando y con el registro de los métodos y estrategias con que lo hace.
Ahora hay una nueva perspectiva: La analítica masiva de datos personalizados. Los algoritmos utilizados en otros medios, adecuadamente orientados por las teorías del aprendizaje personalizado, por técnicas pedagógicas y de diseño instruccional pueden, junto con los avances en minería de datos, obtener informaciones para ajustar mejor la intervención educativa, para mejorar el rendimiento del alumnos, a más de su satisfacción, y el del programa educativo
This paper presents a review of the big data learning analytics in Higher Education. It relates to a new taskbased and achievement-based learning paradigm in line with individual capacities, but not with time, space or age. Its feasibility and relevance are clearly defined by the 2 sigma problem, which raises the broad horizon ahead to a limit target learning. Currently, a strong interest in analysis of learning data is found using systems and software based on social and ubiquitous environments and on the new LMS, which have them included. The problem is that until now, commonly used consolidated tools only get data and graphs that relate individual against group performance, and both together. Moreover, those relationships only make reference to learning data taken from conventional evaluation input. However, there is a space that provides a huge amount of data not only for student evaluation which we currently ignore, at least explicitly. It is the space of connected personal work, networking with peers, with teachers, with the resources and all the material to be used, and with the registration of the methods and strategies they use. Now there is a new perspective: personalised big data analytics. Algorithms used for other media, properly guided by personalized learning theories, by teaching techniques and instructional design can, along with advances in data mining, obtain information to set educational intervention better, to improve student performance, besides their satisfaction and that of the educational program of student orientation or instructional design is essential
This paper presents a review of the big data learning analytics in Higher Education. It relates to a new taskbased and achievement-based learning paradigm in line with individual capacities, but not with time, space or age. Its feasibility and relevance are clearly defined by the 2 sigma problem, which raises the broad horizon ahead to a limit target learning. Currently, a strong interest in analysis of learning data is found using systems and software based on social and ubiquitous environments and on the new LMS, which have them included. The problem is that until now, commonly used consolidated tools only get data and graphs that relate individual against group performance, and both together. Moreover, those relationships only make reference to learning data taken from conventional evaluation input. However, there is a space that provides a huge amount of data not only for student evaluation which we currently ignore, at least explicitly. It is the space of connected personal work, networking with peers, with teachers, with the resources and all the material to be used, and with the registration of the methods and strategies they use. Now there is a new perspective: personalised big data analytics. Algorithms used for other media, properly guided by personalized learning theories, by teaching techniques and instructional design can, along with advances in data mining, obtain information to set educational intervention better, to improve student performance, besides their satisfaction and that of the educational program of student orientation or instructional design is essential
Keywords
Analítica de datos masivos; Educación universitaria; Detección abandono; Analítica de datos masivos; Paradigma aprendizaje postindustrial; Personalización; Problema 2 sigmas; Aprendizaje divergente; College education; Drop detection; Massive data analytics; Postindustrial learning paradigm; Divergent learning; Learning Analytics; Personalization; 2 Sigma Problem







