Criterion-based assessment of chemical concept maps using artificial intelligence

Анотація

This article proposes an artificial intelligence–based criterion-driven model for assessing chemical concept maps in chemistry education. The model is designed to evaluate not only the number of links between concepts, but also their scientific validity, logical coherence, hierarchical organization, and semantic depth. By focusing on the quality of conceptual connections, the proposed approach enables a more objective measurement of students’ conceptual understanding. The model aims to standardize assessment practices, support meaningful formative feedback, and provide teachers with analytical insights into students’ conceptual structures in chemistry. In chemistry education, meaningful learning depends on students’ ability to construct coherent relationships among concepts rather than memorizing isolated facts. Concept maps are widely used to visualize such relationships; however, their assessment often remains subjective and inconsistent, relying heavily on individual teacher judgment. Measuring the depth and correctness of conceptual links poses a particular challenge. The integration of artificial intelligence into criterion-based assessment offers a timely solution by enhancing reliability, transparency, and scalability in evaluating chemical concept maps. This approach aligns with current demands for data-informed, formative, and learner-centered assessment practices in science education [1].