ChemConcept Bridge: An AI-Powered Adaptive Chemistry Learning and Secure Examination System with Misconception Detection, Predictive Analytics, and Dynamic Concept Graph Integration

Chemistry is an essential subject, yet many students
struggle with it due to its abstract concepts, and traditional
teaching methods often emphasize theory over intuitive
understanding, creating a significant “understanding gap” where
misconceptions can easily develop and persist. To address this
issue, we developed ChemConcept Bridge, an AI-based
Intelligent Tutoring System (ITS) designed to enhance
conceptual clarity and engagement. At its core is an adaptive
quiz engine that identifies specific conceptual errors and adjusts
difficulty levels in real time. The system employs predictive
models such as K-Nearest Neighbors (KNN), Naive Bayes,
Decision Trees, Support Vector Machines (SVM), and
Backpropagation Neural Networks to analyze student
performance and behavior, enabling personalized learning
pathways and accurate prediction of student needs. It also
incorporates a secure monitoring tool to maintain academic
integrity in online environments. A key feature, the Dynamic
Chemical Concept Graph (DCCG), acts as a self-evolving
knowledge map that connects chemical concepts dynamically,
allowing targeted interventions and adaptive learning paths
based on interaction data and common areas of difficulty. Initial
results demonstrate that combining predictive analytics,
gamified engagement, and dynamic concept mapping
significantly improves student retention and performance, while
also enabling early identification of struggling learners, allowing
timely teacher intervention; moreover, the system is designed to
be scalable and cost-effective, ensuring continuous learning
support even in the absence of direct supervision.