This study investigated the suitability of different methodological approaches to automatic semantic tagging in the analysis of cultural traits as they emerge from subjective meaning reactions to given words (EMUs). Elicited data from British native speakers were collected and coded manually and with an automatic tagging system (Wmatrix). The results of manual coding were then compared to the results offered by Wmatrix, at different levels and using a variety of methods. Furthermore, automatic tagging was applied to 10,000 sentences extracted from a general Web corpus and containing the node word, and the results of the Web corpus were compared to those of the elicited data. Though further investigation is needed, each of the experiments described provide interesting information for the definition of a method in the use of large corpora for the extraction of EMUs
Understanding culture. Automatic semantic analysis of a general Web corpus and a corpus of elicited data.
BIANCHI, Francesca
2010-01-01
Abstract
This study investigated the suitability of different methodological approaches to automatic semantic tagging in the analysis of cultural traits as they emerge from subjective meaning reactions to given words (EMUs). Elicited data from British native speakers were collected and coded manually and with an automatic tagging system (Wmatrix). The results of manual coding were then compared to the results offered by Wmatrix, at different levels and using a variety of methods. Furthermore, automatic tagging was applied to 10,000 sentences extracted from a general Web corpus and containing the node word, and the results of the Web corpus were compared to those of the elicited data. Though further investigation is needed, each of the experiments described provide interesting information for the definition of a method in the use of large corpora for the extraction of EMUsI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
