Hybrid Logical-Statistical Metadata Analysis for Qualitative and Quantitative Study

The aim of this paper is to present a hybrid logical-statistical approach for metadata analysis from biomedical articles. A key problem in data science is the availability of high-quality data, especially in terms of veracity. The approach proposed develops a situation-based modal logic semantics in which the zones of a scientific article along with their atomic metadata are mapped onto an algebraic space of situations. This semantic framework will subsequently serve as the foundation for a statistical analysis of the extracted metadata. The result is a well-motivated qualitative and quantitative analysis of the descriptive metadata extracted from biomedical articles.