ANALYSIS OF CUSTOMER EXPERIENCE IN THE HOTEL BUSINESS

Abstract

In the process of marketing research, it is necessary to process large volumes of accumulated reviews, so it is obvious that the process of processing and analysis should be automated. Modern technical capabilities allow you to create an information and analytical system for formalized analysis of consumer reviews about hotels. The analysis of consumer reviews contributes to the development of the hotel business, since reviews are a qualitative way to assess the effectiveness of the company. With their help, it becomes clear what elements of the hotel complex should be paid attention to with the aim of their possible improvement in the future. The trends observed in reviews also allow you to find out what consumers expect and how best of all you can live up to their expectations. In fact, the task of analyzing the tonality of the text is equivalent to the problem of text classification, where tonal estimates can be categories of texts. To teach the classification model, a set of marked data downloaded from the Internet was used. The set contains 18,500 reviews about the hotels of Europe in Russian with the tonality indicated for them. The article is devoted to the urgent problem of determining the consumer attitude to services provided in the hotel business based on reviews from the Internet. A solution to the problem of analyzing the tonality of the entities of the hotel, extracted from the reviews about the hotel. To solve the problem, the MLP algorithm is selected, since the task of analyzing tonality is reduced to the task of classification, where tonal estimates are classification classes. Particular attention is paid to the algorithm for teaching a classification model developed using the Keras library in Python. To solve the task of analyzing the tonality of entit