• The Journal of Korean Association of Computer Education
• /
• v.18 no.5
• /
• pp.63-70
• /
• 2015

In the environment that the complex conditions need to be satisfied, skyline query have been applied to various field. To processing a skyline query in centralized scheme, several techniques have been suggested and recently map/reduce platform based approaches has been proposed which divides data space into multiple partitions for the vast volume of multidimensional data. However, the performances of these approaches are fluctuated due to the uneven data loading between servers and redundant tasks. Motivated by these issues, this paper suggests a novel technique called MR-DEAP which solves the uneven data loading using the random sampling. The experimental result gains the proposed MR-DEAP outperforms MR-Angular and MR-BNL scheme.

• Journal of KIISE:Databases
• /
• v.29 no.4
• /
• pp.247-261
• /
• 2002

Aggregation is an important operation that affects the performance of OLAP systems. In this paper we define a new class of aggregation queries, called range-groupby queries, and present a method for processing them. A range-groupby query is defined as a query that, for an arbitrarily specified region of an n-dimensional cube, computes aggregations for each combination of values of the grouping attributes. Range-groupby queries are used very frequently in analyzing information in MOLAP since they allow us to summarize various trends in an arbitrarily specified subregion of the domain space. In MOLAP applications, in order to improve the performance of query processing, a method of maintaining precomputed aggregation results, called the prefix-sum array, is widely used. For the case of range-groupby queries, however, maintaining precomputed aggregation results for each combination of the grouping attributes incurs enormous storage overhead. Here, we propose a fast algorithm that can compute range-groupby queries with minimal storage overhead. Our algorithm maintains only one prefix-sum away and still effectively processes range-groupby queries for all possible combinations of the grouping attributes. Compared with the method that maintains a prefix-sum array for each combination of the grouping attributes in an n-dimensional cube, our algorithm reduces the space overhead by (equation omitted), while accessing a similar number of cells.

• The KIPS Transactions:PartD
• /
• v.11D no.2
• /
• pp.281-292
• /
• 2004

Selectivity estimation for spatial query is very important process used in finding the most efficient execution plan. Many works have been performed to estimate accurate selectivity. Although they deal with some problems such as false-count, multi-count, they can not get such effects in little memory space. Therefore, we propose a new technique called MW Histogram which is able to compress summary data and get reasonable results and has a flexible structure to react dynamic update. Our method is based on two techniques : (a) MinSkew partitioning algorithm which deal with skewed spatial datasets efficiently (b) Wavelet transformation which compression effect is proven. The experimental results showed that the MW Histogram which the buckets and wavelet coefficients ratio is 0.3 is lower relative error than MinSkew Histogram about 5%-20% queries, demonstrates that MW histogram gets a good selectivity in little memory.

• Journal of the Korea Computer Industry Society
• /
• v.4 no.4
• /
• pp.443-448
• /
• 2003

Approximate k-nearest neighbor queries are frequently occurred for finding the k nearest neighbors to a given query point in spatial database systems. The number of searched nodes in an index must be minimized in order to increase the performance of approximate k nearest neighbor queries. In this paper. we suggest the technique of approximate k nearest neighbor queries on R-tree family by improving the existing algorithm and evaluate the performance of the proposed method in dynamic spatial database environments. The simulation results show that a proposed method always has a low number of disk access irrespective of object distribution, size of nearest neighbor queries and approximation rates as compared with an existing method.

• Journal of Korea Spatial Information System Society
• /
• v.6 no.1 s.11
• /
• pp.59-71
• /
• 2004

Nearest neighbor query retrieves nearest located target objects, and is very frequently used in mobile environment. In this paper we propose a novel neatest neighbor query processing technique that is able to retrieve nearest located target object from the user who is continuously moving with a direction. The proposed method retrieves objects using the direction property of moving object as well as euclidean distance to target object. The proposed method is applicable to traffic information system, travel information system, and location-based recommendation system which require retrieving nearest located object.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10b
• /
• pp.4-6
• /
• 2003

본 논문에서는 대규모 DNA 시퀀스를 위한 트라이 인덱싱 기법을 기반으로 하는 효율적인 부분 시퀀스 검색 기법을 제시한다. 제안된 인덱싱 방안에서는 저장 공간 감소를 위하여 시퀀스의 각 문자를 최소 비트 정보로 표현하며, 저장 구조로서 포인터를 사용하지 않는 디스크 기반의 이진 접미어 트라이 구조를 사용한다. 질의 처리 방안에서는 포인터가 없는 이진 트라이 구조 상에서 질의 시퀀스를 검색하기 위하여 이진 정보 기반의 연산과정을 필요로 하며, 또한 단말 정보를 효율적으로 검색하기 위하여 별도의 단말정보 테이블과 인덱스 구조를 사용한다. 실험 결과에 의하면 제안된 방식은 기존의 접미어 트리 인덱싱 방식에 비하여 약 30~50%의 저장 공간 감소 효과를 가질 뿐 아니라, 평균 질의 처리 시간에 있어 약 20배까지의 성능 개선 효과를 갖는 것으로 나타났다.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.396-401
• /
• 2021

초거대 언어모델를 활용한 퓨샷(few shot) 학습법은 여러 자연어 처리 문제에서 좋은 성능을 보였다. 하지만 데이터를 활용한 추가 학습으로 문제를 추론하는 것이 아니라, 이산적인 공간에서 퓨샷 구성을 통해 문제를 정의하는 방식은 성능 향상에 한계가 존재한다. 이를 해결하기 위해 초거대 언어모델의 모수 전체가 아닌 일부를 추가 학습하거나 다른 신경망을 덧붙여 연속적인 공간에서 추론하는 P-tuning과 같은 데이터 기반 추가 학습 방법들이 등장하였다. 본 논문에서는 문맥에 따른 질의 정규화 문제를 대화형 음성 검색 서비스에 맞게 직접 정의하였고, 초거대 언어모델을 P-tuning으로 추가 학습한 경우 퓨샷 학습법 대비 정확도가 상승함을 보였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06a
• /
• pp.381-384
• /
• 2011

최근 무선 센서 네트워크의 수집 데이터에 대해 에너지 효율적인 저장 및 질의 처리를 위한 다양한 연구가 이루어지고 있다. 데이터 중심 저장 (DCS: Data-Centric Storage) 기법은 인-네트워크 방식 기반의 효율적인 데이터 저장과 질의 처리를 위해 제안된 기법이다. DCS 기법은 수집 데이터의 값에 따라 저장 될 위치를 미리 결정하여 각 데이터가 발생시 해당 위치에 인-네트워크 방식으로 저장한다. 이를 통해 질의 처리시 불필요한 질의 배포를 최소화 시킨다. 하지만 기존에 제안된 DCS 기법들은 수집되는 데이터의 발생 범위를 고정적으로 설정한다. 따라서 시기별로 상이한 범위의 데이터가 발생되는 실제 응용에 서는 저장 공간 활용의 불균등을 초래하여 네트워크 수명을 단축시킨다. 본 논문은 시간이 지남에 따라 변화 하는 데이터 발생 패턴에 상황 적응적인 범위 설정 기법을 적용하여 네트워크 전반에 걸쳐 노드들의 저장 공간을 균등하게 사용하는 상황 인지 데이터 중심 저장 방식을 제안한다. 또한 제안하는 기법의 우수성을 보이기 위해 기존 DCS 기법과 성능을 비교평가 한다.

• The KIPS Transactions:PartD
• /
• v.11D no.4
• /
• pp.835-844
• /
• 2004

Recently, the location-based service for moving customers is becoming one of the most important service in mobile communication area. For moving object applications, there are lots of update operations and such update loads are concentrated on some particular area unevenly. Range queries, whose range can be window or circular, are the most essential query types in LBS. We need to distinguish completely contained cells from partially contained cells in those range queries. Also, it is necessary to consider temporal dimension to filter out qualifying objects correctly. In this paper, we adopt two-level index structures with fixed grid file structures in the second level, which are designed to minimize update operations. We propose a spatial ceil filtering method using VP filtering and a combined spatio-temporal filtering method using time gone concepts. Some experimental results are shown for various window queries and circular queries with different filtering combinations to show the performance tradeoffs of the proposed methods.

• Journal of KIISE:Databases
• /
• v.30 no.6
• /
• pp.593-605
• /
• 2003

Since spatial join processing consumes much time, several algorithms have been proposed to improve spatial join performance. The M-way R-tree join (MRJ) is a join algorithm which synchronously traverses M R-trees in the M-way spatial join. In this paper, we introduce indirect predicates which do not directly come from the multi-way join conditions but are indirectly derived from them. By applying the concept of indirect predicates to MRJ, we improve the performance of MRJ. We call such a multi-way R-tree join algorithm using indirect predicates indirect predicate filtering (IPF). Through experiments using synthetic data and real data, we show that IPF significantly