Artificial intelligence applications in education: Natural language processing in detecting misconceptions DOI Creative Commons
Yunus Kökver, Hüseyin Miraç Pektaş, Harun Çelik

и другие.

Education and Information Technologies, Год журнала: 2024, Номер unknown

Опубликована: Авг. 6, 2024

Abstract This study aims to determine the misconceptions of teacher candidates about greenhouse effect concept by using Artificial Intelligence (AI) algorithm instead human experts. The Knowledge Discovery from Data (KDD) process model was preferred in where Analyse, Design, Develop, Implement, Evaluate (ADDIE) instructional design cycle used. dataset obtained 402 analysed Natural Language Processing (NLP) methods. classified Machine Learning (ML), one AI tools, and supervised learning algorithms. It concluded that 175 did not have sufficient knowledge effect. found with highest accuracy rate used predict candidates’ Multilayer Perceptron (MLP). Furthermore, through Enhanced Ensemble Model Architecture developed researchers, combination ML algorithms has achieved rate. kappa (κ) value examined determining significant difference between expert evaluation, it there a difference, strength agreement according research findings. findings current represent alternative prevailing pedagogical approach, which increasingly come rely on information technologies improving conceptual understanding detection misconceptions. In addition, recommendations were made for future studies.

Язык: Английский

Artificial intelligence applications in education: Natural language processing in detecting misconceptions DOI Creative Commons
Yunus Kökver, Hüseyin Miraç Pektaş, Harun Çelik

и другие.

Education and Information Technologies, Год журнала: 2024, Номер unknown

Опубликована: Авг. 6, 2024

Abstract This study aims to determine the misconceptions of teacher candidates about greenhouse effect concept by using Artificial Intelligence (AI) algorithm instead human experts. The Knowledge Discovery from Data (KDD) process model was preferred in where Analyse, Design, Develop, Implement, Evaluate (ADDIE) instructional design cycle used. dataset obtained 402 analysed Natural Language Processing (NLP) methods. classified Machine Learning (ML), one AI tools, and supervised learning algorithms. It concluded that 175 did not have sufficient knowledge effect. found with highest accuracy rate used predict candidates’ Multilayer Perceptron (MLP). Furthermore, through Enhanced Ensemble Model Architecture developed researchers, combination ML algorithms has achieved rate. kappa (κ) value examined determining significant difference between expert evaluation, it there a difference, strength agreement according research findings. findings current represent alternative prevailing pedagogical approach, which increasingly come rely on information technologies improving conceptual understanding detection misconceptions. In addition, recommendations were made for future studies.

Язык: Английский

Процитировано

4