СПЕЦИАЛИЗИРОВАННЫЕ ИНСТРУМЕНТЫ ВИЗУАЛИЗАЦИИ РЕЗУЛЬТАТОВ DATA MINING

Abstract

The relevance of the use of visualization technology in interpreting the results of data analysis is substantiated. Practical aspects of visualization of Data Mining results are considered. Classifications of analytical methods and existing practical tasks are given. The classification of visualization tools in the context of analytical methods and data analytics tasks is proposed. An overview of modern analytical systems allowing to get a visual representation of the data is made. The choice of analytical systems, the visual components of which are used in the study, is justified. A brief description of the tasks of analytics is given. The analysis of the built-in visualizers for various analytical systems is carried out taking into account the methods and tasks of analytics. The similarities and differences in the built-in visualization tools used for each of the systems are considered. The problem of the lack of visual components for interpreting the results of the "associative rules" model is indicated. Modern visualization tools are noted on the example of interactive panels "Cluster Profiles" in the Loginom package. The analysis of specialized tools for visualization of metrics in the considered Data Mining models is carried out. Conclusions are formulated about the possibility of supplementing the built-in tools of a specific analytical system with alternative components of other systems.