SciNeM is a data science tool for metapath-based querying and analysis of Heterogeneous Information Networks (HINs). It currently supports the following operations, given a user-specified metapath: (i) ranking entities using a random walk mode, (ii) retrieving the most similar pairs of entities, (iii) finding the most similar entities to a query entity, and (iv) discovering entity communities via several community detection algorithms. All supported operations have been implemented in a scalable manner, utilising Apache Spark for scaling out through parallel and distributed computation. SciNeM has a modular architecture making it easy to extend it with additional algorithms and functionalities. Moreover, it provides an intuitive, Web-based user interface to build and execute complex constrained metapath-based queries and to explore and visualise the corresponding results.
More details can be found in this publication and in the video presentation below.Back