Ontology Mapping Based on Conditional Information Quantity

Current ontology mapping methods use various mapping strategies to compute similarities between entities, and assign weights to the strategies to combine their results. The formulas for computing similarity and weight should be set according to the characteristics and semantic contexts of strategies. In this paper, the concept of conditional information quantity is proposed. Based on the concept an ontology mapping algorithm is developed, which is able to highlight those more semantically valuable information and filter those noisy or repetitive information. Experiment shows that this algorithm yields results with good precision and recall.