A "Document Contents Representation" (DCR) model is introduced from a formal viewpoint to deal with the entire contents of a document such as individual sentences of a text, bibliography, references, etc. in a scientific information system. A "Mapping Definition Language" (MDL) is proposed to map directly and naturally the document contents into the DCR model. An application of the DCR model and MDL to scientific documents is shown. Some examples of advanced retrieval by SCAT-IR system implemented on the basis of the DCR model and MDL are illustrated.
/lp/association-for-computing-machinery/document-contents-representation-model-of-sentence-retrieval-system-4zChiK1sr1