AUTOMATIC CLASSIFICATION OF EFL LEARNERS’ SELF-REPORTED TEXT DOCUMENTS ALONG AN AFFECTIVE CONTINUUM

Oct 15, 2026

Study proposes an automatic machine-learning approach to position EFL learners along an affective continuum using their self-reported written texts. Written self-reports from 475 students across 5 faculties in 3 universities in Turkey were collected and manually labeled as positive, negative, or neutral. Two dataset combinations (AC-2 and AC-3) with different class counts evaluated classification approaches. Results show strong agreement between automatic and manual labeling and the highest accuracy reaches 90.06% on AC-2 and 71.79% on AC-3, with top discriminative features supporting affective assessment.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Introduction

Affective domain and its components

Affective variables in language learning

Motivation and attitude in second language acquisition

Anxiety and language learning

DOI: 10.20535/2410-8286 .248091 AUTOMATIC CLASSIFICATION OF EFL LEARNERS’ SELF-REPORTED TEXT DOCUMENTS ALONG AN AFFECTIVE CONTINUUM Derya Uysal 1, Alper Kürşat Uysal2 (Corresponding author) 1School of Foreign Languages, Alanya Alaaddin Keykubat University, Alanya, Turkey 2Department of Computer Engineering, Alanya Alaaddin Keykubat University, Alanya, Turkey [alper.uysal@alanya.edu.tr](alper.uysal@alanya.edu.tr) This study aims to place EFL learners along an affective continuum via machine learning methods and present a new dataset about affective characteristics of EFL learners. In line with the purposes, written self-reports of 475 students from 5 different faculties in 3 universities in Turkey were collected and manually assigned by the researchers to one of the labels (positive, negative, or neutral) . Asa result, two combinations ofthe same dataset (AC-2 and AC-3) including different numbers of classes were used for the assessment of automatic classification approaches. Results revealed that automatic classification confirmed the manual classification to a great extent and machine learning methods could be used to classify EFL students along an affective continuum according to their affective characteristics. Maximum accuracy rate of automatic classification is 90.06% on AC-2 dataset including two classes. Similarly, on AC-3 dataset including three classes, maximum accuracy rate of classification is 71.79% . Last, the top-10 features/words obtained by feature selection methods are highly discriminative in terms of assessing student feelings for EFL learning. It could be stated that there is not an existing study in which feature selection methods and classifiers are used in the literature to automatically classify EFL learners’ feelings. Keywords: affective factors; EFL learning; text classification; feature selection; EFL students; higher education; affective barriers. Introduction There is a notion called the affective domain and positive feelings are driving forces to attain language learners’ goal of language acquisition. The idea behind this study is to propose an alternative machine learning-based approach in order to place learners along a continuum based on their affective characteristics as EFL (English as а foreign language) learners instead of qualitative, or quantitative approaches. The notion of the affective domain was initially developed by Krathwohl et al. (1964) and it has received considerable contributions from researchers and experts in the field of education so far (Sousa, 2016) . Affective domain addresses to feelings and emotions, as well as their outward expression. To provide a more operational definition, the affective domain has three subcomponents: behaviour, feeling, and cognition. Feeling is the personal sensation experienced. Cognition is the personal judgements that accompany feelings, and behaviour is the observable reaction which involves both cognition and feelings (Brett, Smith, & Huitt, 2003) . Emotions, on the other hand, involve expressive behaviour, bodily reactions and subjective feelings as a result of one’s interaction with the context (Cahour, 2013) . Affective variables are significant for learning in general (Schutz & Lanehart, 2002) and for language learning in particular (Méndez López & Peña Aguilar, 2013) . Unlike any other subject matters, a level of personal engagement is necessary to learn a foreign language. As recognized by language teachers, learners have to cope with the ambiguities and stress of interaction within the parameters of an unfamiliar culture while conveying conversationally appropriate and personally meaningful messages through unfamiliar phonological, semantic, and syntactic systems. That is why most learners find this process inherently stressful (Horwitz, 1995; Kębłowska, 2012) . Language learning is potentially the most threatening experience of all other school subjects since learners are expected to express themselves using a la

FAQs

What is the main goal of the study?

To automatically classify EFL learners’ self-reported texts along an affective continuum using machine learning methods and a newly used dataset.

How were the learners’ texts prepared and labeled?

Self-reports from 475 students were collected and manually assigned to one of three labels: positive, negative, or neutral.

What were the dataset setups used for evaluation?

Two combinations of the same dataset were used: AC-2 and AC-3, differing in the number of classes.

Oct 15, 2026

Uploaded by Nguyễn Văn Học