Michael Tueller, HYLAS: A new metrical search tool for Greek and Latin poetry
Hylas (hylas.org) is an algorithmic scanning and search tool for Greek and Latin poetry. The paper outlines the development of the tool and its coding in Python, from the rules-based model usually followed by classical scholars to a simplified machine-learning process, which has produced a series of databases by which the program judges each possible choice for syllable combination, prosody, scansion, and word division. At each step, the code judges all possibilities in parallel, assigning and compiling a probability ‘score’ for each before making a final judgement. By this method, the algorithm has achieved 100\% accuracy for two metrical schemes, as measured by more than unique 2,000 spot-checks in a corpus of more than 200,000 lines.
Paolo Monella Curriculum
DH bibliography
Paolo Monella home page