Digital learning environments are now being en-hanced with artificial intelligence (AI) technology. As many educational institutions are utilizing online, hybrid, and tra-ditional classroom formats, the use of AI will continue to grow. In traditional physical classroom settings, educators have the opportunity to assess their student’s level of engagement, based upon the student’s visible behaviour or nonverbal cues, such as attention, participation, and behaviour. However, in online learning environments, where students’ behaviours and nonverbal cues are limited to the instructor’s interaction with the student, it may be difficult for the instructor to monitor their students’ behaviour and to determine which students appear to be disengaged. In order to address these challenges, an Automated Data-Profiling Assistant (ADPA) system is proposed in this research study, which is a data profiling assistant that utilizes artificial intelligence (AI) based multimodal learning analytics to monitor and evaluate student engagement. This system collects informa-tion from various interaction data sources, including: webcam-based facial expressions, keyboard activity, mouse activity, screen interactions, and learning activity logs. Using this data, a data pre-processing module will clean, normalise, and structure the interaction data in a format for analysis. Following this, the ADPA will extract feature types, including behaviour patterns of emotion, response time, typing speed, and navigation behaviour. Long Short Term Memory (LSTM) Networks have been used to model students’ sequential learning behavior and create an effective classification of their engagement into three levels (engaged, moderately engaged or disengaged) so as to allow for accurate identification of time-related changes in student attention that are typically undetectable through traditional machine-learning algorithms. In addition, the system is capable of providing real-time feedback and alerts, which will allow teach-ers to intervene effectively by using personalized instructional techniques and resources. Results from experiments show that integrating multiple sources of data with deep learning models significantly increases the reliability and accuracy of engagement tracking compared to the conventional technique. Therefore, the proposed system provides an innovative and scalable approach to improving student engagement, learning success, and efficiency in today’s educational environments.

The purpose of this research involved creating an artificial intelligence (AI)-based study assistant to track and analyze how students engage in a digital learning environment. Given the rapid increase of online educational platforms, educators around the world are increasingly concerned with monitoring and evaluating student engagement [1][2]. The proposed model utilizes learning analytics along with machine learning (ML) and deep learning (DL) techniques to analyze student interactions with course materials and deter-mine their level of engagement. Various types of engagement-related data such as behavioral logs, facial expressions, and response patterns are collected and processed through modules including data preprocessing, feature extraction, and classifi-cation [3][4]. Among the evaluated models, the Long Short-Term Mem-ory (LSTM) model achieved the best performance with an accuracy rate of 93%, significantly outperforming traditional machine learning approaches [5][6]. This is due to its ability to effectively capture temporal dependencies in sequential data. The proposed study assistant provides several benefits in modern educational environments. It enables educators to identify at-risk or disengaged students, deliver personalized feedback, and implement data-driven instructional strategies. This ultimately leads to improved academic performance, increased student motivation, and reduced dropout rates in online learning systems [7][8]. Future research may focus on the development of transformer-based models, integration of multimodal data sources, and real-time engagement monitoring systems. These advancements will contribute to more intelligent and effective educational technologies, enhancing overall digital learning experiences [9][10].

[1]R. S. Baker and K. Yacef, “The state of educational data mining in 2009,” Journal of Educational Data Mining, vol. 1, no. 1, pp. 3–17, 2009. [2]S. D’Mello and A. Graesser, “AutoTutor and affective autotutor,” ACM Transactions on Interactive Intelligent Systems, vol. 2, no. 4, 2012. [3]J. Whitehill et al., “Faces of engagement: Automatic recognition of student engagement from facial expressions,” IEEE Transactions on Affective Computing, vol. 5, no. 1, pp. 86–98, 2014. [4]Y. Bengio, I. Goodfellow, and A. Courville, Deep Learning. MIT Press, 2016. [5]I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA: MIT Press, 2016. [6]T. Mikolov et al., “Recurrent neural network based language model,” in Proc. INTERSPEECH, 2010. [7]S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997. [8]J. Devlin et al., “BERT: Pre-training of deep bidirectional transformers,” in Proc. NAACL, 2019. [9]A. Krizhevsky et al., “ImageNet classification with deep convolutional neural networks,” in Proc. NIPS, 2012. [10]T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning. Springer, 2009. [11]C. Bishop, Pattern Recognition and Machine Learning. Springer, 2006. [12]H. Peng et al., “Multimodal engagement detection in online learning,” IEEE Access, 2020. [13]L. Zeng et al., “Learning analytics in higher education,” Educational Technology & Society, 2018. [14]P. Dillenbourg, Virtual Learning Environments. Springer, 2015. [15]J. Han et al., Data Mining: Concepts and Techniques. Morgan Kauf-mann, 2011. [16]S. Russell and P. Norvig, Artificial Intelligence: A Modern Approach. Pearson, 2016. [17]R. S. Baker, “Educational data mining and learning analytics,” in Handbook of Learning Analytics, 2017. [18]B. Woolf, Building Intelligent Interactive Tutors. Morgan Kaufmann, 2010. [19]M. Worsley and P. Blikstein, “Multimodal learning analytics,” in Proc. LAK Conference, 2014. [20]G. Siemens and R. Baker, “Learning analytics and educational data mining,” in Proc. LAK, 2012. [21]A. Kapoor et al., “Predicting student engagement,” in Proc. ACM CHI, 2015. [22]K. Koedinger et al., “Learning analytics and data mining,” Educational Psychologist, 2015. [23]D. Romero and S. Ventura, “Educational data mining: A review,” IEEE Transactions on Systems, 2013. [24]S. K. D’Mello, “A selective meta-analysis on emotion in learning,” Educational Psychologist, 2013. [25]Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, pp. 436–444, 2015. [26]A. Graves, Supervised Sequence Labelling with RNNs. Springer, 2012. [27]Z. Zhang et al., “Deep learning for engagement detection,” IEEE Access, 2019. [28]M. Chaudhuri et al., “Student engagement detection in e-learning,” Computers & Education, 2020. [29]L. Chen et al., “Multimodal engagement recognition,” IEEE Transac-tions on Multimedia, 2018. [30]H. Liu et al., “Deep neural networks for learning analytics,” IEEE Access, 2021. [31]J. Brown et al., “AI in education,” Educational Technology Review, 2019. [32]P. Blikstein, “Multimodal learning analytics,” in Proc. LAK, 2017. [33]S. Buckingham Shum and R. Ferguson, “Social learning analytics,” Educational Technology, 2012. [34]A. Kapoor and R. Picard, “Multimodal affect recognition,” in Proc. ACM Multimedia, 2005. [35]T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proc. KDD, 2016.

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close