Computing Poetry

We present SPARSAR, a system for the automatic analysis of poetry(and text) style which makes use of NLP tools like tokenizers, sentence splitters, NER (Name Entity Recognition) tools, and taggers. Our system in addition to the tools listed above which aim at obtaining the same results of quantitative linguistics, adds a number of additional tools for syntactic and semantic structural analysis and prosodic modeling. We use a constituency parser to measure the structure of modifiers in NPs; and a dependency mapping of the previous parse to analyse the verbal complex and determine Polarity and Factuality. Another important component of the system is a phonological parser to account for OOVWs, in the process of grapheme to phoneme conversion of the poem. We also measure the prosody of the poem by associating mean durational values in msecs to each syllable from a database and created an algorithm to account for the evaluation of durational values for any possible syllable structure. Eventually we produce six general indices that allow single poems as well as single poets to be compared. These indices include a Semantic Density Index which computes in a wholly new manner the complexity of a text/poem.