Data abnormal feature detection method and device

The invention discloses a data abnormal feature detection method and device, relates to the technical field of data mining, and mainly aims to improve the automation degree of data abnormal feature detection. According to the main technical scheme, the method comprises the steps of obtaining configuration information input by a user through a configuration interface, wherein the configuration information comprises abnormal thresholds about at least two data sets; extracting common characteristic variables in the at least two data sets; analyzing the distribution states of the variable values corresponding to the characteristic variables in each data set one by one; comparing the difference degrees of the distribution states of the characteristic variables in different data sets; when the difference degree is greater than the abnormal threshold, determining the characteristic variable as an abnormal characteristic; and displaying and outputting the determined abnormal characteristics. 本发明公开了一种数据异常特征的检测方法及装置，涉及