OreChem ChemXSeer: A semantic digital library for chemistry

Representing the semantics of unstructured scientific publications will certainly facilitate access and search and hopefully lead to new discoveries. However, current digital libraries are usually limited to classic flat structured metadata even for scientific publications that potentially contain rich semantic metadata. In addition, how to search the scientific literature of linked semantic metadata is an open problem. We have developed a semantic digital library oreChem Chem_xSeer that models chemistry papers with semantic metadata. It stores and indexes extracted metadata from a chemistry paper repository Chem_xSeer using "compound objects". We use the Open Archives Initiative Object Reuse and Exchange (OAI-ORE))¹ standard to define a compound object that aggregates metadata fields related to a digital object. Aggregated metadata can be managed and retrieved easily as one unit resulting in improved ease-of-use and has the potential to improve the semantic interpretation of shared data. We show how metadata can be extracted from documents and aggregated using OAI-ORE. ORE objects are created on demand; thus, we are able to search for a set of linked metadata with one query. We were also able to model new types of metadata easily. For example, chemists are especially interested in finding information related to experiments in documents. We show how paragraphs containing experiment information in chemistry papers can be extracted and tagged based on a chemistry ontology with 470 classes, and then represented in ORE along with other document-related metadata. Our algorithm uses a classifier with features that are words that are typically only used to describe experiments, such as "apparatus", "prepare", etc. Using a dataset comprised of documents from the Royal Society of Chemistry digital library, we show that the our proposed method performs well in extracting experiment-related paragraphs from chemistry documents.