Method for Evaluation the Pattern of Internet Service Customers Based on Stylometric Analysis Oof their Text Content
Valeriy Sydorenko, Heorhii Chyzhmak, Maryna Oladko · 2022 IEEE 4th International Conference on Modern Electrical and Energy System (MEES) · 2022
Messengers and chats of Internet services (in particular, transport companies, taxi services, energy generating companies etc.) are traditional tools for customer communication with the service provider and a means of feedback. Text messages from customers are a source of operational data and can contain important information about customer preferences, and opinions about certain service elements and quality. Tools of semantic and sentiment analysis allow us to give answers to these questions. However, the authors show the stylometric analysis of chat messages gives a lot of additional information about the personality of the customer himself, who often remains anonymous. The article proposes a method of segmentation of Internet service customers in the space of their stylometric characteristics. The authors also showed that in the space of stylometric characteristics there are certain homogeneous groups that have stable both stylometric patterns. These results can be the basis for the development of a classifier that will allow classifying a new customer by the nature of writing and optimizing his service. The use of such a method by managers or chatbots, in particular, based on artificial intelligence, will increase the level of service of the customers of service of power generating companies, etc.