ScholarWorks@숙명여자대학교

초록

In many scientific fields, array-type data is increasing in size and stored. In particular, data obtained in these fields have a two-dimensional array of numerical values. Scientists demand high-level analysis based on complex mathematical algorithms, rather than simple calculations, from large two-dimensional array data. In this paper, a method of finding outlier regions, whose size is not determined in advance, from large scale scientific two-dimensional array data is researched. The proposed method starts to search for outlier regions with size 1 and expands it one step at the upper, lower, left, and right side of the region. This extension is continued until the extended outlier region no longer meets the outlier region condition, and the maximally extended outlier region is returned to the user as the detected outlier region. Therefore, the proposed method can efficiently search for outlier regions without searching the outlier regions with all possible sizes.

키워드