Кластерный анализ в области предупреждения и ликвидации пожаров и чрезвычайных ситуаций

Abstract

Introduction. At the present stage, the task of high-quality and effective processing of a large amount of information that is collected and accumulated in databases, reference books, documents and other sources of information in the field of fire prevention and liquidation and emergency situations, fire safety is becoming more and more urgent. One of the most popular methods of information analysis is cluster analysis, which allows dividing the set of objects and features under study into homogeneous groups (clusters). As a result of cluster analysis, it is possible to identify patterns of changes in the parameters (values) under study, classify them and transform information into knowledge. Goals and objectives. The purpose of the review is to analyze Russian and foreign scientific publications to identify the main directions and methods of using cluster analysis, special software, development tools and programming languages. Methods. During the review of scientific publications, information was collected from open sources, a list of publications was formed taking into account the accepted selection criteria and search keywords. As a result, the classification of the directions of application of cluster analysis in the field of fire prevention and liquidation and emergency situations, fire safety is carried out. Results and their discussion. The analysis of the reviewed publications shows that cluster analysis methods using modern computer technologies are used in a wide range of research in the field of fire prevention and emergency response. In general, it is possible to classify six main areas of application of cluster analysis in the subject area under consideration. When conducting cluster analysis, in most cases, universal statistical software tools are used, as well as the authors' own developments and high-level programming languages with numerous libraries for multidimensional data analysis. In methodological terms, conducting research on cluster analysis requires a systematic description of all stages of work. This is especially true for the issues of standardization (normalization) of data and evaluation of the quality of clustering. Conclusion. The paper reviews 15 scientific Russian and foreign publications on the application of cluster analysis in the field of fire prevention and emergency response. As a result of the analysis, the main directions of application of cluster analysis in the subject area under consideration, clustering methods, applied tools and programming languages are determined. Proposals are formulated for the further development of the application of cluster analysis, taking into account the need for a more detailed description of the stages of its implementation. Keywords: cluster analysis, fire, emergency, multidimensional data analysis, statistics.