DEVELOPMENT OF ALGORITHMS OF INTELLIGENT SERVICE FOR INFORMATION SEARCH AND MONITORING

Abstract

This paper describes the problem of strategic analysis and the choice of directions for the development of an innovative enterprise in the conditions of transition to the 6th technological order and industry 4.0. In these conditions, search and analytical processing of information cannot be fully performed without the use of automated information and analytical systems, including those based on artificial intelligence. During the analysis, the main priority functions that the developed services should provide were identified. The main difficulties in the development of these services are identified, such as: pre-processing of data and automated checking of the relevance of databases. To effectively solve thetasks set, the intelligent monitoring and information retrieval service should use an integrated approach, taking into account the effectiveness of applying methods for individual subtasks, and ensure high efficiency of implementing all stages of the intelligent monitoring procedure. In this regard, this paper describes not only the development of a general intelligent search algorithm, but also individual block algorithms necessary to ensure the priority functions of the service being developed. The paper presents the following algorithms: an information search algorithm necessary to solve the problem of full-text search of documents within the database of information resources of the information and analytical complex; an algorithm for the procedure for entering new documents; an algorithm for pre-processing data that includes stemming and removing punctuation marks for subsequent text analysis; an algorithm for evaluating the ranking and relevance of information, including vectorization of documents; an algorithm for clustering information search results based on the Kohonen neural network; the algorithm for checking the relevance of information is to check whether the local copy of the document corresponds to the current version on the source's web resource. The Python programming language for the implementation of the presented algorithm is proposed and justified. The system provides automated continuous monitoring with a high frequency of sending a request without the participation of an operator, which will increase the quality and efficiency of information search in conditions of a large volume of unstructured information.