F1000Research · 2026
Machine learning algorithms to predict digital skills in university teaching staff.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Moreira-Choez Jenniffer Sobeida, Núñez-Naranjo Aracelly Fernanda, Carrasco-Valenzuela Asia Cecilia, López-López Héctor Luis, Vázquez Meza Jesús Alejandro, Sabando-García Angel Ramón
Аннотация
BACKGROUND: The digital transformation of higher education has intensified the need to assess and enhance the digital competencies of university faculty. This study analyzed the effectiveness of various machine learning algorithms in predicting levels of faculty digital competence based on socio-educational variables. The objective was to develop an advanced predictive model, applied to faculty members from the State University of Milagro and the Technical University of Manabí.
METHODS: A quantitative approach was adopted, with a cross-sectional correlational design. Digital competencies were measured using the internationally validated DigCompEdu Check-In instrument, structured across six core dimensions. In the predictive phase, nine supervised machine learning algorithms were trained and evaluated: logistic regression, decision trees, random forest, gradient boosting, k-nearest neighbors, support vector machines, stochastic gradient descent, artificial neural networks, and Naive Bayes. The models were trained using a dataset comprising 5,148 observations, and their performance was assessed using standard classification metrics: area under the ROC curve (AUC), accuracy, F1-score, sensitivity, and Matthew's correlation coefficient (MCC).
RESULTS: Gradient boosting, random forest, and neural network models demonstrated superior predictive performance, particularly at advanced competence levels (B2 and C1). Significant associations were identified between academic level, age, gender, and digital competencies. Logistic regression and Naive Bayes showed limitations in identifying low competence levels (A1), while intermediate levels were often overestimated across several models.
CONCLUSIONS: The findings confirm that machine learning algorithms can accurately predict university faculty digital competencies. Advanced models outperformed traditional ones, especially at higher competence levels. It is recommended to incorporate contextual variables and validate the models in diverse educational settings.
METHODS: A quantitative approach was adopted, with a cross-sectional correlational design. Digital competencies were measured using the internationally validated DigCompEdu Check-In instrument, structured across six core dimensions. In the predictive phase, nine supervised machine learning algorithms were trained and evaluated: logistic regression, decision trees, random forest, gradient boosting, k-nearest neighbors, support vector machines, stochastic gradient descent, artificial neural networks, and Naive Bayes. The models were trained using a dataset comprising 5,148 observations, and their performance was assessed using standard classification metrics: area under the ROC curve (AUC), accuracy, F1-score, sensitivity, and Matthew's correlation coefficient (MCC).
RESULTS: Gradient boosting, random forest, and neural network models demonstrated superior predictive performance, particularly at advanced competence levels (B2 and C1). Significant associations were identified between academic level, age, gender, and digital competencies. Logistic regression and Naive Bayes showed limitations in identifying low competence levels (A1), while intermediate levels were often overestimated across several models.
CONCLUSIONS: The findings confirm that machine learning algorithms can accurately predict university faculty digital competencies. Advanced models outperformed traditional ones, especially at higher competence levels. It is recommended to incorporate contextual variables and validate the models in diverse educational settings.