PPR-SSM

PPR-SSM is a software tool for entity linking in biomedical literature. It aims to link entities mentioned in documents to concepts from domain-specific ontologies, such as those related to chemical compounds, phenotypes, gene-product localization, and processes. The tool utilizes Personalized PageRank (PPR) and the relations within the ontology to generate a graph of candidate concepts for the mentioned entities. By leveraging the knowledge encoded in the domain-specific ontology, PPR-SSM calculates the coherence of a set of candidate concepts, thereby improving the accuracy of entity linking compared to other state-of-the-art methods. Additionally, the tool explores weighting the edges between candidate concepts using semantic similarity measures (SSM) to enhance its performance further. PPR-SSM is a graph-based method that does not require training data, making it more adaptable to various biomedical domains. The tool has demonstrated significant improvements in entity linking accuracy, particularly for chemical compounds, compared to methods that do not use SSMs.