Visualizing Mappings Between Pairwise Ontologies - An Empirical Study of Matrix and Linked Indented List in Their User Support During Class Mapping Creation and Evaluation

Visual support designed to facilitate human interaction with semantic data has largely focused on visualizing entities such as classes and relationships within an ontology. Comparatively speaking, less attention has focused on visualizing mappings established between independent ontologies and determining the effectiveness of mapping visualizations. This paper presents a user study of the matrix and the linked indented list visualization in their visual support for users during creation and evaluation of class mappings between pairwise ontologies. A total of 81 participants took part in a task-based controlled experiment, with the aim of assessing the extent to which a given visualization supports recognition of visual cues, validation of existingmappings, and creation of new results. Based on empirical evidence collected from the participants in their speed to complete the tasks, their success in answering various questions, as well as their physiological sensory data such as eye gaze, we aim to quantify user performance and visual attention demanded in the use of the two aforementioned mapping visualizations. The experimental results indicate that the linked indented lists and the matrix visualization are comparable in terms of effectiveness and efficiency when assisting users in the given task scenarios with marginal differences. However, linked indented lists are likely to demand less effort from the users' visual perceptual systems with statistically significant differences found in several gaze measures, including the physical distances needed to locate relevant visual information, number of fixations and time required to process visual cues, and the overall efforts in scanning the visual scene.