Educational Data Mining & Learning Analytics
Educational Data Mining (EDM) and Learning Analytics (LA) use data analysis to enhance education by extracting patterns and trends from student performance, behavior, and interactions. These disciplines optimize learning processes, identify struggling students, and personalize learning experiences. Integrating them promises to revolutionize education and foster student success in the digital age.
Current DaSCI research lines:
- Portability of predictive models. In educational data mining, a key challenge is ensuring the portability of predictive models across different courses. Currently, the excessive reliance on low-level attributes during training limits the models’ adaptability. To address this issue, we are actively developing and utilizing high-level attributes with richer semantic meaning, aiming to enhance the models’ transferability.
- Fusion of multimodal data. The student data at present is multimodal, sourced from various channels such as learning technology, behavior, physiology, psychometrics, and learning environments. Our approach involves creating predictive models that integrate this diverse information, offering a comprehensive understanding of students’ complete learning journey. The ultimate goal is to improve the precision of predicting student dropout rates by gaining unprecedented insights into their minute-by-minute activities.
- Trustworthy and fair EDM models. In education, trustworthy and fair predictive models are vital. With transparency and bias mitigation, these models gain confidence and adoption from educators. They ensure equal opportunities for diverse students, fostering inclusivity and supporting better learning outcomes. Emphasizing these principles cultivates a data-driven educational environment where students’ individual needs are met, leading to enhanced academic success and personal growth.
|Gibaja Galindo, Eva Lucrecia||egibaja@uco.W.3Gpd3es||–|
|Romero Morales, Cristobal||cromero@ucClvv7Ho.es||–|