Application of text mining techniques to identify problem solving patterns in applied statistics

Author

León Contreras, Víctor  

Abstract

An interesting perspective to evaluate open activities is to track and analyze the actions of the students during their resolution. To do this, the dashboards can provide an overview of the intensity of work, the choices made by the students in solving the activity and the time needed, as well as the estimation of a grade. This would be the case of the dashboard developed by ASISTEMBE which, together with R Commander Traced, allows obtaining and analyzing the traces of the students while they carry out a statistics activity. This modified version of I Commander allows to obtain all the records of the students during the activity, that is, the sequence of actions they perform. The dashboard visually analyzes part of the information contained in the records and represents the performance of the students throughout the activity. In this project, an analysis of these records is performed to detect patterns of similarity between students and identify common actions. Furthermore, the student records have been represented as a sequence of events using the main function of I contained in each of the actions. The visualizations developed for the resolution of the work have been integrated into the dashboard. This allows you to get an overview of student performance and possible resolution patterns without considering any predefined solution path.

 

Director

Gallemí Rovira, Oriol 
Berzosa Rodríguez, Xavier

Degree

IQS SE - Undergraduate Program in Industrial Engineering

Date

2020-05-29