The Semantic Data Dictionary – An Approach for Describing and Annotating Data

17 tables and two figures of this paper. Table 1 is a subset of explicit entries identified in NHANES demographics data. Table 2 is a subset of implicit entries identified in NHANES demographics data. Table 3 is a subset of NHANES demographic Codebook entries. Table 4 presents a subset of explicit entries identified in SEER. Table 5 is a subset of Dictionary Mapping for the MIMIC-III Admission table. Table 6 shows high-level comparison of semantic data dictionaries, traditional data dictionaries, approaches involving mapping languages, and general data integration tools. Table A1 shows namespace prefixes and IRIs for relevant ontologies. Table B1 shows infosheet specification. Table B2 shows infosheet metadata supplement. Table B3 shows dictionary mapping specification. Table B4 is a codebook specification. Table B5 is a timeline specification. Table B6 is properties specification. Table C1 shows NHANES demographics infosheet. Table C2 shows NHANES demographic implicit entries. Table C3 shows NHANES demographic explicit entries. Table C4 presents expanded NHANES demographic Codebook entries. Figure 1 is a conceptual diagram of the Dictionary Mapping that allows for a representation model that aligns with existing scientific ontologies. The Dictionary Mapping is used to create a semantic representation of data columns. Each box, along with the “Relation” label, corresponds to a column in the Dictionary Mapping table. Blue rounded boxes correspond to columns that contain resource URIs, while white boxes refer to entities that are generated on a per-row/column basis. The actual cell value in concrete columns is, if there is no Codebook for the column, mapped to the “has value” object of the column object, which is generally either an attribute or an entity. Figure 2 presents (a) A conceptual diagram of the Codebook, which can be used to assign ontology classes to categorical concepts. Unlike other mapping approaches, the use of the Codebook allows for the annotation of cell values, rather than just columns. (b) A conceptual diagram of the Timeline, which can be used to represent complex time associated concepts, such as time intervals.