Towards a technology for actantial pattern recognition

Abstract

A semiotic model able to recognize occurrences we use during our conversations on social network would be implemented to refine the existing sentiment analysis tools and evolve the pattern recognition technologies. The purpose of this article is introducing a semiotic method taking over a narrative structure of the interactions on Facebook. Starting from the analysis of conversations as enunciative acts, the method is able to turn them into a specific narrative scheme. With this approach we can select and organize the correct linguistic elements to be mined by semantic engines or text mining technologies, to evolve the dialogue between semiotics, data scientists and IT specialists.