Automated Analysis of Product Representation in Digital Advertising of Educational Services

Abstract

The article examines the structural and semantic characteristics of the representation of goods in digital advertising of educational services of helping specialists in social networks (using the example of the social network Vkontakte). The possibilities of automated analysis methods in identifying verbal markers of representations of professional tools for further creation of monitoring systems and ethical evaluation of information about electronic goods are being investigated. A general scheme has been developed for the automated collection of information about educational products and processing of the data obtained by thematic communities (teachers, psychologists, doctors). The LDA (Latent Dirichlet Allocation) method was used for thematic modeling of text data. Using the TF-IDF (Term Frequency-Inverse Document Frequency) method, important terms were identified in each set of descriptions for analyzing semantic frames of advertisements. Verbal markers have been identified, based on which the authors of advertising texts introduce the nodal elements of the professional training purchase scenario for a specialist into the reader's mind. The cluster structure of representations of pedagogical goods includes highly specialized descriptions intended for specialists working in general education institutions; representations of psychological goods are characterized by orientation to a wide audience, including non-specialists, and a thematically mixed structure; in the representations of medical products, the main information is not related to medicine, which can mislead users. The methodology of automated cluster analysis of marketing strategies selling educational content will be in demand by a wide range of specialists in the field of education management, content developers of educational Internet channels, educational institutions, teachers working in a virtual environment.