Increasing evidence from novel research demonstrates that circRNAs play a role in the onset and
development of various diseases. Recently, numerous computational methods have been
proposed to predict circRNA-disease associations, as examining such associations merely by
experimenting is too expensive and time-consuming to be feasible.
In this work, we first discuss the theoretical foundations of searching for circRNA-disease
associations. We then propose a method to construct a graph of circRNAs, diseases and other
types of RNAs and map its entities to vectors, so as to use the said vectors to correctly label
circRNA-disease pairs as either 'related' or 'unrelated'. Afterward, we compare our labels to the
ground truth and evaluate our model in terms of prevailing evaluation metrics.