Divide and Conquer Technique for Large Linked Datasets
Maria Krommyda, Verena Kantere, Yannis Vassiliou · 2019
Significant effort has been dedicated in recent years to the exploitation of very large linked datasets due to the importance of the information they contain and the increase of their availability. Some techniques have been developed that handle the volume of these datasets by aggregating their information based on the data structure or model. Other approaches exploit specific characteristics of the datasets, such as semantic annotations, to present the information to the users in semantically defined ways. In an era that the volume and diversity of the available information increases exponentially and more users are interested in exploring it, it is crucial to provide a technique that will allow the exploitation of diverse and very large linked datasets in a scalable way independent of any characteristics of the input dataset. We present here a generic divide and conquer technique that can offer the required scalability to exploit any input dataset regardless its size and characteristics. The proposed technique has been tested in the context of interactive representation of very large linked datasets as graphs.