Enhancing Dataset Discovery With Knowledge Graph Link Prediction Techniques

● In the evolving landscape of open science, the ability to navigate and discover pertinent datasets is increasingly significant. This primarily hinges on the presence of detailed metadata, delineating the dataset’s content, and potential spheres of application.

● The GES DISC datasets are characterized by science keywords to enable dataset discovery in web search interfaces.

● A problem may arise where a dataset lacks a science keyword that it otherwise should have.

● Machine learning techniques such as link prediction can be used to detect these missing science keywords by estimating the probability of new links forming between dataset and keyword nodes.