Knowledge tracing model based on human-machine collaboration: an analysis of the impact of perceptual ambiguity, selective attention, and heuristic judgment on learning performance

In different phases of learning, such as perception, attention, inspiration, and reasoning, the importance of educators’ professional guidance often remains underestimated, despite its profound impact on instructional outcomes. This study introduces a new approach called Human-Machine Collaboration-based Knowledge Tracing (HMCKT), highlighting the crucial role of human-machine collaboration within the model, which differs from previous versions of Knowledge Tracing (KT) models. This study method, named Teach-Study Active Learning (TSAL), strategically selects the most informative samples for annotation, mimicking the dynamic interaction between educators and learners, and delineating effective knowledge concepts (KCs). Additionally, the Spatio-Temporal Graph Convolutional Network (STGCN) facilitates the training of knowledge states across various time intervals and spatial contexts, reflecting the complex learning trajectory of learners. Ultimately, this integration yields a robust knowledge framework. By conducting empirical research, this study clarifies how factors such as perceptual ambiguity, selective attention, and heuristic judgment influence the learning process. These findings provide valuable insights for improving learning strategies and pedagogical approaches.