Machine Learning for Processing Customer Feedback Texts in Marketing Tasks

In modern times, the capabilities of machine learning are rapidly increasing in all areas of human activity, including marketing, without the use and processing of textual information and images. Machine learning has significantly eased the process of working with visual content. Big companies are trying to target and personalize their products. They do this by analyzing people's interests and pulling them in the right direction. It is a proven method that helps organizations to attract a specific audience. In order to maximize the use of investments, a correct orientation to the buyer is required. Without analyzing the audience's wishes in fact, one risks significant losses and customer distrust. The paper presents the use of the clustering algorithm to detect similarities in the text. Problems of context-dependent text analysis based on grammar and feature extraction based on n-grams are presented. The tasks of keyword phrase extraction and total gist detection are discussed. The paper presents the possibilities of using clustering algorithms in marketing, in particular sales tasks, which allows people with similar characteristics to be grouped according to their interests in certain products. At the same time, the clustering algorithm will group products based on customer feedback. As a result, a picture of how much the product is in demand is obtained, which will be taken into account in order to improve sales. Thus, in sales tasks, processing customer feedback text for subsequent content-analyse is very effective in obtaining product information.