During the last decade, we witnessed an exponential rise of datasets from heterogeneous sources. Ontologies are playing an essential role in consistently describing domain concepts, data harmonization and integration to support large-scale integrative analysis and semantic interoperability in knowledge sharing. Several semantic similarity (SS) measures have been suggested to enable the integration of rich ontology structures into automated reasoning and inference. However, there is no tool that exhaustively implements these measures and existing tools are generally Gene Ontology specific, do not implement several models suggested in the WordNet context and are not equipped to properly deal with frequent ontology updates. We introduce a Python SS measure library (PySML), which tackles issues related to current SS tools, providing a portable and expandable tool to a broad computational audience. This empowers users to manipulate SS scores from several applications for any ontology version and file format. PySML is a flexible tool enabling the implementation of all existing semantic similarity models, resolving issues related to computation, reproducibility and re-usability of SS scores.
|Number of pages||17|
|Publication status||Published - 18 Aug 2021|