Development of a technique for automatic detection signs of suggestion in Russian advertising texts based on rules and neural network technologies

In the modern world, the field of marketing is actively developing, the amount of online advertising is increasing every year. According to statistics, in 2023 the growth of the online advertising market was approximately 30% compared to the previous year, and in 2024 growth is expected to be 19-26%. Such a rapid increase in the amount of advertising is associated primarily with its effectiveness, which is achieved through the use of various manipulative and suggestive techniques. Advertising writers use suggestions at different levels of language to influence consumer behavior. Linguists often have to analyze texts and detect suggestion manually. However, identifying means of suggestion and determining the degree of suggestibility of a text require significant human resources. Modern methods of computational linguistics, such as natural language processing, machine learning methods and deep learning methods, can solve this problem. That is why the development of a method for automatic detection of suggestion is an urgent task. The purpose of the study is to develop a method for automatically detecting suggestions in Russian-language advertising. The following tasks follow from the set goal: to study existing automatic methods for detecting suggestive signs, to develop a method for detecting suggestion, to automatically collect and process an array of advertising texts, to compare automatic methods for extracting each suggestive sign and to choose the best method, to obtain the significance of each of the suggestive signs based on expert assessments , obtain a metric for assessing the suggestiveness of advertising text, create a Web interface for extracting signs of suggestion and determining the suggestiveness of the text, carry out, according to the selected classification of means of suggestion, a linguistic analysis of errors in automatic detection of suggestion.