I. Introduction
The problem of detecting anomalies, especially outliers, is a highly relevant issue. Due to the significant increase in the analyzed data, it is of great importance to indicate effective methods of detecting outliers. Anomalies in the data can occur through system malfunctions, or human error. Regardless of the source of the error, the literature provides a number of methods for detecting outliers in the data. A common feature of a number of outlier detection methods is the use of distance between data [1], in particular using the nearest neighborhood [2]–[5] or data distribution [6], [7]. Recently, there are more and more works using fuzzy techniques to detect anomalies [8]–[17]. In addition, a very interesting approach is to use information granules to detect outliers [17], [18].