A MODELLING OF ADVERTISEMENT TEXT INTENTIONS ON A BASIS OF FRAME SYSTEM

Abstract

The article considers a model of perception of advertisement text intentions. Advertisement text contains a set of intentions, which are part of a speech genre of advertising. The model is based on Fillmore's frames; however, drawbacks of this approach are also described. Disadvantages include ambiguity and mixture of markers. To avoid these drawbacks, the denotation graph is used to describe the structure of advertisement text. Any denotatum can be expanded to a full frame if it is necessary, as both denotatum and frame are mental constructs. The method is tested. The article gives an explication of denotation structure and a visualization of denotation graph. A possible training set and a classifier architecture (artificial neural network) are provided for the method application.