TaLTaC2 (T2) is a ‘metric’ software for the automatic analysis of texts of various types (collections of documents and fragments, short texts from surveys, websites etc.). T2 allows for both a lexical and a textual treatment, and for the full integration of (unstructured) textual data with (structured) codified data - the latter associated with the units of context. The lexical unit of the analysis is mixed, both words and multiwords. The search and extraction of information exploits statistico-linguistic resources – both standard and tailor-made – as well as weighing criteria of a statistical and quantitative nature. Linguistic and semantic annotations are also employed. This allows the user to individuate keywords and elements of terminology at the level of the whole corpus, of parts of it and of single units of context. The results can be exported via matrices, and this makes multi-dimensional analyses possible that express associations between words and latent sense models.
Taltac, una piattaforma per l’analisi automatica dei testi / Bolasco, Sergio. - ELETTRONICO. - (2011). (Intervento presentato al convegno Ricerca qualitativa e giovani studiosi tenutosi a Enna nel 1-3 settembre 2011).
Taltac, una piattaforma per l’analisi automatica dei testi
BOLASCO, Sergio
2011
Abstract
TaLTaC2 (T2) is a ‘metric’ software for the automatic analysis of texts of various types (collections of documents and fragments, short texts from surveys, websites etc.). T2 allows for both a lexical and a textual treatment, and for the full integration of (unstructured) textual data with (structured) codified data - the latter associated with the units of context. The lexical unit of the analysis is mixed, both words and multiwords. The search and extraction of information exploits statistico-linguistic resources – both standard and tailor-made – as well as weighing criteria of a statistical and quantitative nature. Linguistic and semantic annotations are also employed. This allows the user to individuate keywords and elements of terminology at the level of the whole corpus, of parts of it and of single units of context. The results can be exported via matrices, and this makes multi-dimensional analyses possible that express associations between words and latent sense models.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.