Verification of Genuineness of Educational Internet Resources Using Machine Learning Methods

The analysis of various sources has been carried out, confirming the demand for protective solutions against information threats from fake websites of organizations. The relevance of the research topic is confirmed by the presence of addresses similar in spelling to popular educational Internet resources in the list of domain names of the RU zone registered over the past year. When developing security solutions, it is proposed to use a site verifying the genuineness method based on the use of machine learning algorithms. A review of research by various authors has been conducted in order to determine the most frequently used features for verifying the genuineness of Internet resources in the machine learning models. The authors propose their own verifying the genuineness features that can be used in a situation where data about an Internet resource is extracted from the domain name registry. The proposed features differ from the known ones in that they contain signs of micro-markup on the site and storage of site settings files for search engines on servers. The developed features were tested using a naive Bayesian classifier as a machine learning algorithm. The proposed features can be used in the implementation of a software tool to counteraction threats from fake educational sites.