Computing and Information Systems - Research Publications
Permanent URI for this collection
Search Results
1 - 8 of 8
-
ItemThe V*Diagram: A query-dependent approach to moving KNN queries(Association for Computing Machinery (ACM), 2008-01-01)The moving k nearest neighbor (M k NN) query finds the k nearest neighbors of a moving query point continuously. The high potential of reducing the query processing cost as well as the large spectrum of associated applications have attracted considerable attention to this query type from the database community. This paper presents an incremental safe-region-based technique for answering M k NN queries, called the V*-Diagram. In general, a safe region is a set of points where the query point can move without changing the query answer. Traditional safe-region approaches compute a safe region based on the data objects but independent of the query location. Our approach exploits the current knowledge of the query point and the search space in addition to the data objects. As a result, the V*-Diagram has much smaller IO and computation costs than existing methods. The experimental results show that the V*-Diagram outperforms the best existing technique by two orders of magnitude.
-
ItemThe hvtree: A memory hierarchy aware version index(Association for Computing Machinery (ACM), 2010-01-01)The huge amount of temporal data generated from many important applications call for a highly efficient and scalable version index. The TSB-tree has the potential of large scalability due to its unique feature of progressive migration of data to larger mediums. However, its traditional design optimized for two levels of the memory hierarchy (the main memory and the hard disk) undermines its potential for high efficiency in face of today's advances in hardware, especially CPU/cache speed and memory size. We propose a novel version index structure called the HV-tree. Different from all previous version index structures, the HV-tree has nodes of different sizes, each optimized for a level of the memory hierarchy. As data migrates to different levels of the memory hierarchy, the HV-tree will adjust the node size automatically to exploit the best performance of all levels of the memory hierarchy. Moreover, the HV-tree has a unique chain mechanism to maximally keep recent data in higher levels of the memory hierarchy. As a result, HV-tree is several times faster than the TSB-tree for point queries (query with single key and single time value), and up to 1000 times faster than the TSB-tree for key-range and time-range queries.
-
ItemGeneralized multidimensional data mapping and query processing(ASSOC COMPUTING MACHINERY, 2005-09)Multidimensional data points can be mapped to one-dimensional space to exploit single dimensional indexing structures such as the B + -tree. In this article we present a Generalized structure for data Mapping and query Processing (GiMP), which supports extensible mapping methods and query processing. GiMP can be easily customized to behave like many competent indexing mechanisms for multi-dimensional indexing, such as the UB-Tree, the Pyramid technique, the iMinMax, and the iDistance. Besides being an extendible indexing structure, GiMP also serves as a framework to study the characteristics of the mapping and hence the efficiency of the indexing scheme. Specifically, we introduce a metric called mapping redundancy to characterize the efficiency of a mapping method in terms of disk page accesses and analyze its behavior for point, range and kNN queries. We also address the fundamental problem of whether an efficient mapping exists and how to define such a mapping for a given data set.
-
ItemiDistance:: An adaptive B+-tree based indexing method for nearest neighbor search(ASSOC COMPUTING MACHINERY, 2005-06)In this article, we present an efficient B + -tree based indexing method, called iDistance, for K-nearest neighbor (KNN) search in a high-dimensional metric space. iDistance partitions the data based on a space- or data-partitioning strategy, and selects a reference point for each partition. The data points in each partition are transformed into a single dimensional value based on their similarity with respect to the reference point. This allows the points to be indexed using a B + -tree structure and KNN search to be performed using one-dimensional range search. The choice of partition and reference points adapts the index structure to the data distribution.We conducted extensive experiments to evaluate the iDistance technique, and report results demonstrating its effectiveness. We also present a cost model for iDistance KNN search, which can be exploited in query optimization.
-
ItemOptimized algorithms for predictive range and KNN queries on moving objects(PERGAMON-ELSEVIER SCIENCE LTD, 2010-12)
-
Item
-
ItemStreaming multiple aggregations using phantoms(SPRINGER, 2010-08)
-
ItemAnalysis and evaluation of V*-kNN: an efficient algorithm for moving kNN queries(SPRINGER, 2010-06)