STUDENT PERFORMANCE PREDICTION USING MACHINE LEARNING ALGORITHMS DOI
Manish Tiwari, Nilesh Jain

ShodhKosh Journal of Visual and Performing Arts, Год журнала: 2024, Номер 5(6)

Опубликована: Июнь 30, 2024

The accurate prediction of student performance is a critical component in enhancing educational outcomes, enabling timely interventions, and personalizing learning experiences. This research paper investigates the application various machine algorithms to predict performance, addressing limitations traditional methods that often fail handle large datasets multiple variables effectively. By leveraging data from academic records, attendance, socio-economic factors, this study evaluates efficacy decision trees, random forests, support vector machines, neural networks identifying at-risk students. methodology includes preprocessing, model training, rigorous evaluation using metrics such as accuracy, precision, recall, F1 score. Cross-validation techniques ensure robustness predictive models. findings reveal models, particularly forests networks, significantly outperform accuracy. Key factors influencing success, including attendance background, are identified, providing actionable insights for educators policymakers. contributes field mining by offering comprehensive analysis applications education proposing robust practical implementation. implications highlight potential revolutionize practices data-driven decision-making fostering an environment conducive success. Future directions include biases exploring integration additional sources further enhance

Язык: Английский

A self‐reported instrument to measure and foster students' science connection to life with the CARE‐KNOW‐DO model and open schooling for sustainability DOI Creative Commons
Alexandra Okada

Journal of Research in Science Teaching, Год журнала: 2024, Номер unknown

Опубликована: Июнь 9, 2024

Abstract National governments are concerned about the disconnection of young people from science, which hampers development a scientifically literate society promoting sustainable development, wellbeing, equity, and green economy. Introduced in 2015 alongside Agenda 2030, “open schooling” approach aims at enhancing students' science connections through real‐life problem solving with families scientists, necessitating solid evidence for scalability sustainability. This study conceptualizes “science connection,” term yet underexplored, as integration science's meaning purpose into personal, social, global actions informed by socioscientific thinking. It details novel 32‐item self‐report questionnaire developed validated insights 85 teachers connection”‐enhanced learning. A new consensual qualitative analysis method visual textual snapshots enabled developing quantitative measures findings rigor. The multilanguage instrument provided just‐in‐time actionable data, immediacy applicability feedback to 2082 underserved students aged 11–18 across five countries participating open schooling activities using CARE‐KNOW‐DO model. innovative feature supports responsible research, offering real‐time fostering immediate educational impact. Exploratory confirmatory factor analyses revealed components connection: Confidence aspiration science; Fun participatory teachers, family, experts; Active learning approaches; Involvement in‐and‐outside school activities; Valuing role life‐and‐society. Many felt connected science— Brazil: 80%, Spain: 79%, Romania: 73%, Greece: 70%, UK: 57%— boys: 75%, girls: nonbinary students: 56%. These differences need in‐depth research. Results suggest that decline primary secondary education, but model may reengage older students. robust connection enhances scientific literacy builds capital. aids policymakers, educators, learners identifying factors facilitate or impede engagement efforts.

Язык: Английский

Процитировано

5
Predictive insights into U.S. students’ mathematics performance on PISA 2022 using ensemble tree-based machine learning models DOI
Li Zhu, Hye Sun You, Minju Hong

и другие.

International Journal of Educational Research, Год журнала: 2025, Номер 130, С. 102537 - 102537

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
A machine‐learning model of academic resilience in the times of the COVID‐19 pandemic: Evidence drawn from 79 countries/economies in the PISA 2022 mathematics study DOI
Kwok‐cheung Cheung, Pou‐seong Sit, Jia‐qi Zheng

и другие.

British Journal of Educational Psychology, Год журнала: 2024, Номер 94(4), С. 1224 - 1244

Опубликована: Сен. 22, 2024

Abstract Background Given that students from socio‐economically disadvantaged family backgrounds are more likely to suffer low academic performance, there is an interest in identifying features of resilience, which may mitigate the relationship between socio‐economic status and performance. Aims This study sought combine machine learning explainable artificial intelligence (XAI) technique identify key resilience mathematics during COVID‐19. Materials Methods Based on PISA 2022 data 79 countries/economies, random forest model coupled with Shapley additive explanations (SHAP) value not only uncovered but also examined contributions each feature. Results Findings indicated 35 were identified classification academically resilient non‐academically students, largely validated previous framework. Notably, gender differences shown distribution some features. Research findings tended have a stable emotional state, high levels self‐efficacy, truancy positive future aspirations. Discussion has established research paradigm essentially methodological nature bridge gap psychological theories big field educational psychology. Conclusion To sum up, our shed light issues education equity quality global perspective times COVID‐19 pandemic.

Язык: Английский

Процитировано

0
STUDENT PERFORMANCE PREDICTION USING MACHINE LEARNING ALGORITHMS DOI
Manish Tiwari, Nilesh Jain

ShodhKosh Journal of Visual and Performing Arts, Год журнала: 2024, Номер 5(6)

Опубликована: Июнь 30, 2024

The accurate prediction of student performance is a critical component in enhancing educational outcomes, enabling timely interventions, and personalizing learning experiences. This research paper investigates the application various machine algorithms to predict performance, addressing limitations traditional methods that often fail handle large datasets multiple variables effectively. By leveraging data from academic records, attendance, socio-economic factors, this study evaluates efficacy decision trees, random forests, support vector machines, neural networks identifying at-risk students. methodology includes preprocessing, model training, rigorous evaluation using metrics such as accuracy, precision, recall, F1 score. Cross-validation techniques ensure robustness predictive models. findings reveal models, particularly forests networks, significantly outperform accuracy. Key factors influencing success, including attendance background, are identified, providing actionable insights for educators policymakers. contributes field mining by offering comprehensive analysis applications education proposing robust practical implementation. implications highlight potential revolutionize practices data-driven decision-making fostering an environment conducive success. Future directions include biases exploring integration additional sources further enhance

Язык: Английский

Процитировано

0