Intelligent visualization techniques for reusable learning objects to facilitate an online learning environment

Abstract

The idea of learning objects and their working standardization for the Learning Object Metadata as recommended by IEEE 1484 provides a conceptual and implementation platform to promote reuse of learning resources with the latest educational technology. These significant changes undoubtedly affect the design of e-learning systems where learning resources like course or quiz materials can be conveniently anchored in a network of well-defined learning objects. Many real-world learning object repositories like the MERLOT consist of thousands or even tens of thousands of potentially interrelated learning objects. To allow the designers of those e-learning systems to restructure the network of learning objects for better reuse and re-purpose existing resources, greater accessibility or faster navigation, and more importantly facilitating both educators and learners to quickly identify some important relation(s) among the involved learning objects (or concepts), adaptive information visualization techniques can help. Through adaptive visualization, users can focus on various subsets of learning objects with interesting properties for careful analysis. In this paper, we have enhanced the well-known force scan algorithm (FSA) for integration with effective heuristics to produce appropriate diagrams of different scales or shapes for visualizing the relations among various learning objects. For rare cases where only independent learning objects are involved, our goal is to improve the visibility by evenly spreading out the learning objects as nodes with adjustable angular displacement while avoiding node overlapping on the different levels. However, for practical diagrams of related learning objects, the main challenge is to avoid both node and edge overlapping while spreading out the concerned learning objects on pre-assigned levels. In both cases, our adaptive visualization algorithms work with their best efforts to preserve the mental map of the initial diagrams for learning objects. We implemented prototypes of the two adaptive visualization algorithms in C++, and evaluated their performance on both random and real test cases whenever available. The experimental results revealed the strength of our proposal from which e-learning systems can benefit greatly. More importantly, the results shed light on several interesting directions for further investigation.