Abstract
In Linked Data, owl: sameAs predicate is used for linking different datasets. However, there are reservations about its use and its role in linking data. In linking different datasets, owl: sameAs can lead to violation of the strict identity semantics of this predicate, because it is symmetric and transitive, meaning that anyone can link their data with your dataset, from anywhere. Consequently, the growth of the sameAs network raises concerns as to how accurate different datasets will be in the future. Therefore, there is a need for more sophisticated applications and systems built upon different published datasets that verify and optimize the accuracy of the Linked Datasets. To achieve that goal, we purpose a method that provides a similarity-based numeric identification score, which reviews different constraints and measuring metrics, and verifies the accuracy of the use of sameAs in the selected datasets. The system utilizes different matching algorithms on datasets comprising of three features, URI data management, statistical data and similarity matching score. The functionalities developed in this system were compared with its closest equivalents. The results obtained from the system indicate that not all claimed subjects linked via owl: sameAs across different entities are similar.