Knowledge Discovery in Spatial Databases

Abstract

Abstract. Both, the number and the size of spatial databases, such as geographic or medical databases, are rapidly growing because of the large amount of data obtained from satellite images, computer tomography or other scientific equipment. Knowledge discovery in databases (KDD) is the process of discovering valid, novel and potentially useful patterns from large databases. Typical tasks for knowledge discovery in spatial databases include clustering, characterization and trend detection. The major difference between knowledge discovery in relational databases and in spatial databases is that attributes of the neighbors of some object of interest may have an influence on the object itself. Therefore, spatial knowledge discovery algorithms heavily depend on the efficient processing of neighborhood relations since the neighbors of many objects have to be investigated in a single run of a typical algorithm. Thus, providing general concepts for neighborhood relations as well as an efficient implementation of these concepts will allow a tight integeration of spatial knowledge discovery algorithms with a spatial database management system. This will speed-up both, the development and the execution of spatial KDD algorithms. For this purpose, we define a small set of database primitives, and we demonstrate that typical spatial KDD algorithms are well supported by the proposed database primitives. By implementing the database primitives on top of a commercial database management system, we show the effectiveness and efficiency of our approach, experimentally as well as analytically. The paper concludes by outlining some interesting issues for future research in the emerging field of knowledge discovery in spatial databases.

Spatial neighborhood based anomaly detection in sensor datasets

“…Specifically we study outlier detection in the neighborhoods based on spatial relationships and (a) JC, (b) SC and (c) JC (AND/OR) SC C. Comparison with other approaches: In order to discuss the improvements that accounting for heterogeneity will bring about in our approach, we first lay out some results with traditional approaches which give us a context for the results we obtained in our approach. Specifically, we discuss the results for -Comparison of our approach for neighborhood formation with traditional cardinality based approach (Ester et al 1999;Shekhar et al 2001). -Comparison of our approach with Graph based spatial anomaly detection technique (Shekhar et al 2001).…”

“…Thus, identifying the neighborhood based only on the spatial relationships would result in one big neighborhood. If we consider cardinality based neighborhood (Ester et al 1999;Shekhar et al 2001) we can have several possible neighborhood formations, since it is not order invariant. Moreover determination of cardinality is based on user's choice similar to our coefficient threshold choice.…”

Anomaly detection in injection molding process data based on unsupervised learning

“…To calculate a clustering decision map for the given set of feature vectors, we will need to make a choice on a clustering algorithm. Within this article, we chose the DBSCAN algorithm by Ester et al [28]. In contrast to other clustering algorithms, DBSCAN is capable of identifying noisy points and, most importantly, finding clusters of non-convex shape, as we are not able to make statements about the shape of the clusters we expect to find.…”

Analysis of delays structure of interconnections in supercomputer by means of DBScan and divisive clustering algorithms

“…Peter Kriegel in 1996 [6]. In this algorithm it is assumed that in space there are «condensations» of objects that form a cluster to each other.…”