http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
GISAR : 공간 분할 지수를 이용한 이미지 데이터 연관 규칙 마이닝 Association Rules Mining of Image data using Spatial Factor
金庚昶,宋任英 弘益大學校 科學技術硏究所 2005 科學技術硏究論文集 Vol.16 No.-
The task of spatial data mining is to discover suggestive and potentially useful information, that is not previously known, from a large scale spatial database. Spatial association rule is a rule that describes the close relationship that exists between one or more spatial objects with other spatial object within the spatial database. In this paper, the frequent item sets extracted from the original image by Max occur, an existing multimedia association rule mining algorithm, are used to find the spatial relationship between frequent item sets. An Image association rules mining method is proposed that uses spatial factor (SF) based on grid cells. In addition, an efficient mining algorithm is proposed that applies the minimum spatial support to the repeatedly occuring items and through the spatial relationship between items. The speed and the performance of the mining results decreased since we considered only the spatial association relationship between frequent item sets extracted by the Max occur algorithm, while removing all infrequent items as outliers. However, since the objective of this paper was to find the spatial relationship between frequent objects in an image by looking into the spatial relationship between frequent item sets, we believe that the discovery of association rules between data through spatial relationship between data objects in an image could result in useful and important knowledge.