Citation matching details

Usage. Please provide citations separated by newlines and the service will show the referenced documents. The documents are searched for in PMC Open Access Subset database retrieved on 01/16/2014.

The service returns a result containing the following fields for each citation:

• citation-text: raw citation text you have provided.
• extracted-metadata: metadata fields the service was able to extract from the citation string. Note that this is not always extracted perfectly; however, it does not have to be since the quality of this data is usually sufficient for the whole algorithm to work correctly.
• matched-documents: list of metadata records of documents from the database that match a given citation along with normalised confidence score (between 0.0 and 1.0).

Machine-readable service. The service provides a RESTful API which can be accessed by external client services.

General description of the algorithm. A citation text is analysed using stochastic parser to extract the metadata information. Extracted fields are compared with records in document database using fuzzy string similarity metrics. The final score used to assess documents most likely referenced by a given citation is computed using machine learning techniques (for details, see our paper arXiv:1303.6906).