Assessing reading fluency in elementary grades: A machine learning approach

This study compares eleven widely used machine learning algorithms to identify the most accurate and comprehensive method for assessing children's reading fluency. Our fluency framework integrates three key dimensions: accuracy in word decoding, reading speed (words per minute), and prosody, wh...

Full description

Saved in:
Bibliographic Details
Main Authors: Gabriel Candido da Silva, Rodrigo Lins Rodrigues, Américo N. Amorim, Lieny Jeon, Emilia X.S. Albuquerque, Vanessa C. Silva, Vinícius F. da Silva, André L.A. Pinheiro, João P.J.R. Nunes, Suzana X.M.G. de Souza, Maxsuel S. Silva, Igor Mauro, Alexandre Magno Andrade Maciel
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:Computers and Education: Artificial Intelligence
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666920X25000517
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849329775857893376
author Gabriel Candido da Silva
Rodrigo Lins Rodrigues
Américo N. Amorim
Lieny Jeon
Emilia X.S. Albuquerque
Vanessa C. Silva
Vinícius F. da Silva
André L.A. Pinheiro
João P.J.R. Nunes
Suzana X.M.G. de Souza
Maxsuel S. Silva
Igor Mauro
Alexandre Magno Andrade Maciel
author_facet Gabriel Candido da Silva
Rodrigo Lins Rodrigues
Américo N. Amorim
Lieny Jeon
Emilia X.S. Albuquerque
Vanessa C. Silva
Vinícius F. da Silva
André L.A. Pinheiro
João P.J.R. Nunes
Suzana X.M.G. de Souza
Maxsuel S. Silva
Igor Mauro
Alexandre Magno Andrade Maciel
author_sort Gabriel Candido da Silva
collection DOAJ
description This study compares eleven widely used machine learning algorithms to identify the most accurate and comprehensive method for assessing children's reading fluency. Our fluency framework integrates three key dimensions: accuracy in word decoding, reading speed (words per minute), and prosody, which captures appropriate pausing and intonation during reading. Audio recordings from 2nd and 3rd grade students across 144 Brazilian schools were transcribed using the advanced Whisper ASR system, enabling automated extraction of fluency features. The research objective was to determine which algorithm best predicts fluency, considering diverse evaluation setups including binary classification (fluent versus non-fluent), multiclass classification (differentiated fluency levels), and regression analysis to estimate continuous fluency scores. Among the eleven models evaluated, Logistic Regression achieved the highest overall performance in the classification experiments, demonstrating superior precision and accuracy. Ensemble methods such as Gradient Boosting and Random Forest also yielded robust results, particularly in regression analyses where they effectively captured the variability in expert fluency ratings. Notably, reading speed emerged as the most critical indicator of fluency across all experiments, consistently outweighing contributions from accuracy and prosody. Nevertheless, prosody maintained significant importance, especially when fluency was modeled as a continuous variable, emphasizing the value of expressive reading. These findings suggest that while high accuracy in word decoding is fundamental, interventions that enhance reading speed and incorporate prosodic features can offer more nuanced support for reading development. This work provides educators with insights into scalable, automated, and cost-effective approaches to monitoring and improving reading fluency.
format Article
id doaj-art-12d6ca9acee14064814d13f054e8de8c
institution Kabale University
issn 2666-920X
language English
publishDate 2025-06-01
publisher Elsevier
record_format Article
series Computers and Education: Artificial Intelligence
spelling doaj-art-12d6ca9acee14064814d13f054e8de8c2025-08-20T03:47:10ZengElsevierComputers and Education: Artificial Intelligence2666-920X2025-06-01810041110.1016/j.caeai.2025.100411Assessing reading fluency in elementary grades: A machine learning approachGabriel Candido da Silva0Rodrigo Lins Rodrigues1Américo N. Amorim2Lieny Jeon3Emilia X.S. Albuquerque4Vanessa C. Silva5Vinícius F. da Silva6André L.A. Pinheiro7João P.J.R. Nunes8Suzana X.M.G. de Souza9Maxsuel S. Silva10Igor Mauro11Alexandre Magno Andrade Maciel12Programa de Pós-Graduação em Engenharia da Computação, Universidade de Pernambuco, Rua Benfica, 455, Recife, 50720-001, PE, Brazil; Corresponding author.Programa de Pós-Graduação em Ensino das Ciências, Universidade Federal Rural de Pernambuco, Avenida Dom Manoel de Medeiros, s/n, Recife, 52171-900, PE, BrazilCenter for Faculty Advancement, New York University, 82 Washington Square E, New York, 10003, NY, United States; Escribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, Brazil; Inteligência Artificial na Educação e Aprendizado (IAEA), CESAR School, Avenida Cais do Apolo, 77, Recife, 50030-220, PE, BrazilSchool of Education and Human Development, University of Virginia, 405 Emmet St S, Charlottesville, VA 22904, United StatesEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilEscribo Leitura, Escribo - Educação do Seu Jeito, Estrada do Arraial, 2501, Recife, 52051-380, PE, BrazilPrograma de Pós-Graduação em Engenharia da Computação, Universidade de Pernambuco, Rua Benfica, 455, Recife, 50720-001, PE, BrazilThis study compares eleven widely used machine learning algorithms to identify the most accurate and comprehensive method for assessing children's reading fluency. Our fluency framework integrates three key dimensions: accuracy in word decoding, reading speed (words per minute), and prosody, which captures appropriate pausing and intonation during reading. Audio recordings from 2nd and 3rd grade students across 144 Brazilian schools were transcribed using the advanced Whisper ASR system, enabling automated extraction of fluency features. The research objective was to determine which algorithm best predicts fluency, considering diverse evaluation setups including binary classification (fluent versus non-fluent), multiclass classification (differentiated fluency levels), and regression analysis to estimate continuous fluency scores. Among the eleven models evaluated, Logistic Regression achieved the highest overall performance in the classification experiments, demonstrating superior precision and accuracy. Ensemble methods such as Gradient Boosting and Random Forest also yielded robust results, particularly in regression analyses where they effectively captured the variability in expert fluency ratings. Notably, reading speed emerged as the most critical indicator of fluency across all experiments, consistently outweighing contributions from accuracy and prosody. Nevertheless, prosody maintained significant importance, especially when fluency was modeled as a continuous variable, emphasizing the value of expressive reading. These findings suggest that while high accuracy in word decoding is fundamental, interventions that enhance reading speed and incorporate prosodic features can offer more nuanced support for reading development. This work provides educators with insights into scalable, automated, and cost-effective approaches to monitoring and improving reading fluency.http://www.sciencedirect.com/science/article/pii/S2666920X25000517Reading fluencyAutomatic speech recognitionLarge-scale assessmentNatural language processing
spellingShingle Gabriel Candido da Silva
Rodrigo Lins Rodrigues
Américo N. Amorim
Lieny Jeon
Emilia X.S. Albuquerque
Vanessa C. Silva
Vinícius F. da Silva
André L.A. Pinheiro
João P.J.R. Nunes
Suzana X.M.G. de Souza
Maxsuel S. Silva
Igor Mauro
Alexandre Magno Andrade Maciel
Assessing reading fluency in elementary grades: A machine learning approach
Computers and Education: Artificial Intelligence
Reading fluency
Automatic speech recognition
Large-scale assessment
Natural language processing
title Assessing reading fluency in elementary grades: A machine learning approach
title_full Assessing reading fluency in elementary grades: A machine learning approach
title_fullStr Assessing reading fluency in elementary grades: A machine learning approach
title_full_unstemmed Assessing reading fluency in elementary grades: A machine learning approach
title_short Assessing reading fluency in elementary grades: A machine learning approach
title_sort assessing reading fluency in elementary grades a machine learning approach
topic Reading fluency
Automatic speech recognition
Large-scale assessment
Natural language processing
url http://www.sciencedirect.com/science/article/pii/S2666920X25000517
work_keys_str_mv AT gabrielcandidodasilva assessingreadingfluencyinelementarygradesamachinelearningapproach
AT rodrigolinsrodrigues assessingreadingfluencyinelementarygradesamachinelearningapproach
AT americonamorim assessingreadingfluencyinelementarygradesamachinelearningapproach
AT lienyjeon assessingreadingfluencyinelementarygradesamachinelearningapproach
AT emiliaxsalbuquerque assessingreadingfluencyinelementarygradesamachinelearningapproach
AT vanessacsilva assessingreadingfluencyinelementarygradesamachinelearningapproach
AT viniciusfdasilva assessingreadingfluencyinelementarygradesamachinelearningapproach
AT andrelapinheiro assessingreadingfluencyinelementarygradesamachinelearningapproach
AT joaopjrnunes assessingreadingfluencyinelementarygradesamachinelearningapproach
AT suzanaxmgdesouza assessingreadingfluencyinelementarygradesamachinelearningapproach
AT maxsuelssilva assessingreadingfluencyinelementarygradesamachinelearningapproach
AT igormauro assessingreadingfluencyinelementarygradesamachinelearningapproach
AT alexandremagnoandrademaciel assessingreadingfluencyinelementarygradesamachinelearningapproach