Analysis and clarification of anomalies and outliers classifications in economic data

In the time of informatization data mining is used in almost all spheres of human activity. The reliability of data mining results directly determines the quality of internal and external processes, reduces the possibility of unforeseen situations, and allows creating accurate models of processes and making realistic predictions. In this regard, the detection and neutralization of deviations in the original data becomes particularly relevant. This work is devoted to the study of existing classifications of anomalies and outliers in economic data. To date, the scientific base for the study of intellectual analysis of economic data is extremely limited. The paper considers various classification approaches to the anomalous elements in the data, provides examples of emissions in economic data and determines the importance of timely detection and elimination of emissions to obtain reliable results. The aim of the study is to analyze the existing theoretical base for data mining and assess the possibility of its application to economic data. As a result of the study, classification features were identified and the existing classifications were grouped. The analysis of works on the subject under study also allows complementing the scientific base with a new classification group.