What does a 23 thousand bigram word cloud look like? An honours thesis.

The FAIMS Team is proud to announce the completion of two honours theses by our extremely dedicated students: Georgia Burnett and Corinne Softley.

Georgia’s thesis, A SEMANTIC SPACE FOR ARCHAEOLOGY: TOWARDS DESCRIPTIVE ONTOLOGIES FOR ARCHAEOLOGY has the following word art as its cover (generated with processing2 and wordcram).

> > ## Abstract > > > > Archaeology suffers from a disconnect between datasets. Without interoperable data, projects remain local, relevant only to a pocket of discussion. Interoperable data is only achievable through standardisation. Metadata has been favoured in archaeology as a means of capturing datasets; however, metadata cannot facilitate the required level of interoperability needed when performing large-­‐scale automated comparisons. Content standards---specifically, standardised vocabularies to ensure semantic interoperability---are needed. Ontologies are the best representation for standardised vocabularies in archaeology. Ontologies not only quantify the term in a hierarchy, but also allow for the term to exist over several layers of the network. > > > > Not only does archaeology require content standards, however; the standards that are designed must reflect the natural language of the discipline. They must reflect the nuanced vocabularies of archaeologists in practice. Text analysis provides a viable way of producing a natural language ontology. This thesis makes the first steps towards a natural language ontology by determining relationships between related terms. Positional indexing, a derivative of latent semantic analysis, provides a method of determining relationships between terms in a domain-­‐specific corpus of literature. By doing so, it ensures an ontology reflective of the domain. Synonym extraction for ontology development is considered a key feature of future research. > >

Dialogue & Discussion