Визначення релевантності інформації, отриманої від пошуково-довідкового сервісу на веб-платформі.

Abstract

The article analyzes some possibilities of a quick determination of the relevance of information obtained from some search and reference services located on various web platforms. An important feature to ensure the survivability of the information on the Internet for reference and search services has been identified. This feature consists in providing services to an array of sample data to verify information obtained from primary sources or through the mechanism of following links, which confirm the information or provide an answer to the request in case of failure. It is proposed to consider a search service or some reference platforms using the production rules of if-THEN structures, depending on the available facts. An indicator of the effectiveness of the formalized information is derived through a determinant of the level of the information error when processing it under the conditions of uncertainty, providing that the determinant highlights the data that is not enough to describe the object of the re-quest, and that there may be a solution to system problems among the data. Its representation through some estimated values is given. When developing a search or reference service on a web platform using the proposed approach, such services will hold not just arrays of information structured by topic, but formed arrays of information based on symbolic transformations using both general and separate schemes for displaying information on the request, as well as with the ability to reformulate tasks and queries with the aim of getting the most complete and different information on the request. A mecha-nism for detecting truth and falsity when receiving responses to the request from a service according to the rules of fuzzy logic, when understanding the set can be replaced by understanding the characteristic function, is proposed. An example of an algorithmic implementation is provided as well. All stated above can be used to optimize reference and search services on web platforms, improve the work of search engines and various distributed information processing services.