Peculiarities of using open data in preparing marketing and commercial management decisions

Abstract

The article examines the features of the use of data from open sources of information, special approaches and techniques in the system of marketing research of enterprises. Improving approaches to integration into modern marketing information systems of modern content analysis tools have been developed and tested, which organically provides constructive synergy of applied methods and means of neuro-fuzzy modeling and clustering of information arrays, statistical analysis of information units including potential and actual consumers. The modern methodological tools of machine learning were tested on the basis of the author’s approaches to detecting "fakes" (unreliable information) in the formation of situational awareness of company management, identifying trends in target markets, analytical processing of changes associated with high risks and uncertainties for business. It is shown that the great variability of the modern information environment (data, content) creates significant prerequisites and significant combinatorial opportunities for generating distorted information in different ways, as well as the dissemination of the latter. Possibilities for detecting inaccurate information (fakes) in the information field of a particular product market have been worked out. Comparison of the results of different models based on the confusion matrix showed that two of the four learning models, namely the neural network and the "random forest" model, did well enough to assess the reliability ("fake") messages. The recommendations on the organization of the formation of source data from open sources of information, improving the quality and reliability of their processing and successful integration of marketing and commercial analytics systems have been constructively developed.