摘要 |
<P>PROBLEM TO BE SOLVED: To provide an outlier detection device, etc. that support or execute detection of outlier within practical time without applying parameter tuning work to a non-linear data set. <P>SOLUTION: An outlier detection device 1 converts each data included in a data set into a bit string with respect to each dimension, and builds an observation area of the data set on the basis of the bit string. Then, the outlier detection device 1 determines data of interest, one by one, from the data included in the data set, and calculates a deviation degree of the data of interest on the basis of data density around the data of interest when removing an area corresponding to the data of interest from the observation area. <P>COPYRIGHT: (C)2013,JPO&INPIT |