Publication: Análisis docente con Machine Learning
Loading...
Official URL
Full text at PDC
Publication Date
2019
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
La presente memoria documenta el proceso de desarrollo de una aplicación de Software para la limpieza y posterior análisis de datos relativos al uso del Campus Virtual Integrado de la Universidad Complutense de Madrid por parte de sus alumnos, mediante algoritmos de minería de datos y aprendizaje automático. Los datos analizados corresponden a ficheros de trazas que contienen información sobre las acciones realizadas en la plataforma, tales como: el usuario que ha realizado la acción, la fecha y hora en la que se ha realizado, y los componentes a los que se han accedido (apuntes, foros, wikis, etc.).
Para ello, se desarrolló un script para el procesamiento y limpieza de datos de entrada, se evaluaron varias librerías de algoritmos de aprendizaje automático, de las cuales se eligió la herramienta Weka, y se desarrolló una aplicación para integrar los dos componentes anteriores, junto con una interfaz de usuario para la ejecución de los algoritmos.
Para la implementación de algoritmos de aprendizaje automático se utilizó la librería Waikato Environment for Knowledge Analysis (WEKA). La aplicación se desarrolló en Java, y la interfaz fue implementada con la biblioteca gráfica Swing. El script de limpieza de datos se desarrolló en Python.
La aplicación desarrollada como parte de este trabajo permite al usuario procesar y analizar los datos de entrada mediante su interfaz gráfica. El usuario puede elegir el fichero a analizar, el algoritmo que se va a aplicar sobre este fichero, y los distintos parámetros para filtrar los datos de entrada. Como resultado, la aplicación genera y muestra por pantalla una gráfica correspondiente a los datos analizados y filtrados.
This text documents the process of development of a Software Application for the processing and further analysis of data relative to the use of the integrated online platform “Campus Virtual” of Universidad Complutense de Madrid by its students, via the application of data mining and machine learning algorithms. The data analyzed represents log files that contain information about the actions performed in this platform, such as: the user that performed it, the date and time and the components that were accessed (notes, forums, wikis, etc.). In order to do this, a script was made for the processing and cleansing of input data, several machine learning libraries were evaluated, of which the one chosen was the Weka suite, and an application was built to integrate these two components, along with a user interface to run these algorithms. For the implementation of machine learning algorithm, the used library was Waikato’s Weka (Waikato Environment for Knowledge Analysis). The application was written in Java, and its interface was implemented using Swing, a widget toolkit for Java. The script for data cleansing was written in Python. The developed application as part of this project allows the user process and analyze the input data via its graphic interface. The user can choose the file to analyze, the algorithm that’s going to be applied, and a series of parameters to filter the input data. As a result, the application generates and displays a graph with the analyzed and filtered data.
This text documents the process of development of a Software Application for the processing and further analysis of data relative to the use of the integrated online platform “Campus Virtual” of Universidad Complutense de Madrid by its students, via the application of data mining and machine learning algorithms. The data analyzed represents log files that contain information about the actions performed in this platform, such as: the user that performed it, the date and time and the components that were accessed (notes, forums, wikis, etc.). In order to do this, a script was made for the processing and cleansing of input data, several machine learning libraries were evaluated, of which the one chosen was the Weka suite, and an application was built to integrate these two components, along with a user interface to run these algorithms. For the implementation of machine learning algorithm, the used library was Waikato’s Weka (Waikato Environment for Knowledge Analysis). The application was written in Java, and its interface was implemented using Swing, a widget toolkit for Java. The script for data cleansing was written in Python. The developed application as part of this project allows the user process and analyze the input data via its graphic interface. The user can choose the file to analyze, the algorithm that’s going to be applied, and a series of parameters to filter the input data. As a result, the application generates and displays a graph with the analyzed and filtered data.
Description
Trabajo de Fin de Grado, Universidad Complutense, Facultad de Informática, Departamento de Arquitectura de Computadores y Automática, Curso 2018/2019