To What Extent Can Text Classification Help with Making Inferences About Students’ Understanding

A. J. Beaumont*, T. Al-Shaghdari

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference publication

Abstract

In this paper we apply supervised machine learning algorithms to
automatically classify the text of students' reflective learning journals from an introductory Java programming module with the aim of identifying students who need help with their understanding of the topic they are reflecting on. Such a system could alert teaching staff to students who may need an intervention to support their learning.

Several different classifier algorithms have been validated on the training data set to find the best model in two situations; with equal cost for a positive or negative classification and with cost sensitive classification. Methods were used to identify those individual parameters which maximise the performance of each algorithm. Precision, recall and F1-score, as well as confusion matrices were used to understand the behaviour of each classifier and choose the one with the best performance.

The classifiers that obtained the best results from the validation were then evaluated on a testing data set containing different data to that used for training.

We believe that although the results could be improved with further work, our initial results show that machine learning could be applied to students' reflective writing to assist staff in identifying those students who are struggling to understand the topic.
Original languageEnglish
Title of host publicationMachine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings
EditorsGiuseppe Nicosia, Panos Pardalos, Renato Umeton, Giovanni Giuffrida, Vincenzo Sciacca
PublisherSpringer
Pages372-383
Number of pages12
ISBN (Electronic)978-3-030-37599-7
ISBN (Print)9783030375980
DOIs
Publication statusPublished - 3 Jan 2020
Event5th International Conference on Machine Learning, Optimization, and Data Science, LOD 2019 - Siena, Italy
Duration: 10 Sep 201913 Sep 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11943 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference5th International Conference on Machine Learning, Optimization, and Data Science, LOD 2019
CountryItaly
CitySiena
Period10/09/1913/09/19

Fingerprint Dive into the research topics of 'To What Extent Can Text Classification Help with Making Inferences About Students’ Understanding'. Together they form a unique fingerprint.

  • Research Output

    Reflective journals: can they be used to identify struggling students?

    Wong, S. H. S., Thompson, E. & Beaumont, A., 2016.

    Research output: Contribution to conferencePaper

  • Cite this

    Beaumont, A. J., & Al-Shaghdari, T. (2020). To What Extent Can Text Classification Help with Making Inferences About Students’ Understanding. In G. Nicosia, P. Pardalos, R. Umeton, G. Giuffrida, & V. Sciacca (Eds.), Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings (pp. 372-383). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11943 LNCS). Springer. https://doi.org/10.1007/978-3-030-37599-7_31