• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.174-177
• /
• 2015

자가 지식 학습 프레임워크는 자연어 텍스트에서 지식 트리플을 생성하기 위한 방법 중 하나로, 문장의 의존 관계 트리 상에서 주어 개체와 목적어 개체 사이의 관계를 패턴으로 학습해 이 패턴을 바탕으로 새로운 지식 트리플을 생성한다. 그러나 이 방법은 의존 관계 트리를 생성하는 도구의 성능에 영향을 받을 뿐만 아니라 생성된 지식 트리플을 반복적으로 사용하는 자가 지식 학습의 특성상 오류가 누적될 가능성이 있다. 이러한 문제점을 해결하기 위해서 본 논문에서는 자가 지식 학습 프레임워크에서 생성된 지식 트리플을 TransR 신뢰도 함수를 사용해 신뢰도 값을 측정하여 그 값에 따라 지식 트리플을 필터링하는 방법을 제안한다. 실험 결과에 따르면 필터링 된 지식 트리플들이 그렇지 않은 지식 트리플들에 비하여 더 높은 정확률을 보여주어, 제안한 방법이 자가 지식 학습 프레임워크의 정확률 향상에 효과적임을 증명하였다.

• Annual Conference of KIPS
• /
• 2008.05a
• /
• pp.227-230
• /
• 2008

본 연구에서는 대용량 궤적 데이터베이스에서 영역 질의를 효과적으로 처리하기 위한 인덱싱 기법에 대하여 논의한다. 먼저, 기존 인덱싱 기법의 문제점을 지적하고, 이러한 문제점을 해결하는 새로운 기법을 제안한다. 제안된 기법에서는 우선 시간 차원을 다수의 시간 구간으로 분할하고, 인덱싱의 대상이 되는 전체 라인 세그먼트들을 시간 구간별로 구분한다. 각 시간 구간에 속하는 라인 세그먼트들에 대하여 별도의 인덱스를 구축한다. 또한, 디스크에서 관리되는 과거 시간 구간에 대한 인덱스들과는 달리 최근 시간 구간에 대한 인덱스는 메인 메모리상에 관리함으로써 삽입과 검색의 성능을 크게 개선할 수 있다. 각 시간 구간에 속하는 라인 세그먼트들은 다음과 같은 방식으로 인덱스를 구축한다. 먼저, 2D-트리를 이용하여 전체 공간 차원을 유사한 수의 라인 세그먼트들이 배정되도록 다수의 셀들로 분할한다. 또한, 분할된 각 셀마다 시공간 차원 (x, y, t)에 대한 별도의 3차원 $R^*$-트리를 두어 보다 상세한 인덱싱을 지원한다. 실험 결과에 의하면, 기존 기법에 비하여 작은 인덱스 구조를 갖으면서도 검색 성능면에서 $300%{\sim}1000%$까지의 성능 향상 효과를 갖는 것으로 나타났다.

• The KIPS Transactions:PartD
• /
• v.13D no.7 s.110
• /
• pp.867-878
• /
• 2006

In the past few years, the research of sensor networks is forced dramatically. Specially, while the research for maintaining the power of a sensor is focused, we are also concerned nth query processing related with the optimization of multiple continuous queries for decreasing in unnecessary energy consumption of sensor networks. We present the fragmentation algorithm to solve the redundancy problem in multiple continuous queries that increases in the count or the amount of transmitting data in sensor networks. The fragmentation algorithm splits one query into more than two queries using the query index (QR-4ree) in order to reduce the redundant query region between a newly created query and the existing queries. The R*-tree should be reorganized to the QR-tree right to the structure suggested. In the result, we preserve 20 percentage of the total energy in the sensor networks.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.2
• /
• pp.213-219
• /
• 2020

In this paper, we propose a classification analysis method for diagnosing and predicting problematic smartphone use in order to provide policy data on problematic smartphone use, which is getting worse year after year. Attempts have been made to identify key variables that affect the study. For this purpose, the classification rates of Decision Tree, Random Forest, and Support Vector Machine among machine learning analysis methods, which are artificial intelligence methods, were compared. The data were from 25,465 people who responded to the '2018 Problematic Smartphone Use Survey' provided by the Korea Information Society Agency and analyzed using the R statistical package (ver. 3.6.2). As a result, the three classification techniques showed similar classification rates, and there was no problem of overfitting the model. The classification rate of the Support Vector Machine was the highest among the three classification methods, followed by Decision Tree and Random Forest. The top three variables affecting the classification rate among smartphone use types were Life Service type, Information Seeking type, and Leisure Activity Seeking type.

• Journal of Advanced Navigation Technology
• /
• v.28 no.3
• /
• pp.315-322
• /
• 2024

In this paper, we proposes real-time operating system based PCM encoder for telemetry system that must transmit frames within a set time. In the case of large aircraft, the complexity of the system is increasing because a lot of state information is measured from each sensor and peripheral device. In addition, as the amount measurement data increases, the role of PCM encoder to transmit frames within a set time is becoming important. Existing encoder is inflexible when changing specifications or implementing additional features. Therefore, a design is needed to supplement this. We propose a PCM encoder design applying uC/OS-II. In order to confirm the validity, a simulation was performed to measure the execution time of the task to confirm the performance.

• Journal of the Korea Society of Computer and Information
• /
• v.10 no.4 s.36
• /
• pp.65-75
• /
• 2005

Moving-objects databases should efficiently support database queries that refer to the trajectories and positions of continuously moving objects. To improve the performance of these queries. an efficient indexing scheme for continuously moving objects is required. To my knowledge, range queries on current positions cannot be handled by the 3D R-tree and the TB-tree. In order to handle range queries on current and past positions. I modified the original 3D R-tree to keep the now tags. Most of spatio-temporal index structures suffer from the fact that they cannot efficiently process range queries past positions of moving objects. To address this issue. we propose an access method, called the Tagged Adaptive 3DR-tree (or just TA3DR-tree), which is based on the original 3D R-tree method. The results of our extensive experiments show that the Tagged Adaptive 3DR-tree outperforms the original 3D R-tree and the TB-tree typically by a big margin.

• Journal of KIISE:Databases
• /
• v.27 no.2
• /
• pp.238-246
• /
• 2000

The distance join has been introduced previously, which finds nearest pairs in the order of distance incrementally among two spatial data sets built with multidimensional indexes like R-trees. We propose efficient K-distance joins when the number(K) of pairs to find is preset. Especially, we develop a distance join algorithm with bi-directional expansion and optimized plane sweeping using selection method of sweep axis and direction. The experiments on real spatial data sets show that the proposed algorithm is much better than the former algorithms.

• Parasites, Hosts and Diseases
• /
• v.34 no.4
• /
• pp.259-266
• /
• 1996

The taxonomic validity of morphological group III Accnthamoeba app. is uncertain. In the present study. six type strains of group III Aconthamoeba spry. , A. culbertsoni, A. heniyi, A. pustulosc, A. palestinensis, A. royrebn and A. lenticulnto were subjected for the evaluation or their taxonomic validity by comparison of the isoeneyme patterns by isoelectic focusing on polyacrylamide gels, mitochondrial DNA (Mt DNA) restriction fragment length polymorphism (RFLP) . and small subunit ribosomal DNA (ssu rDNA) PCR-RFLP patterns. The Mt DNA RFLP patterns were heterogeneous between the species. The type strains of A. pclestinensls and A. pustulosc showed almost identical patterns of isoenrymes and rDNA PCR-RFLP with an estimated sequence divergence of 2.6%. The other species showed heterogeneous patterns of isoenxymes and rDNA PCR- RFLP. It is likely that A. pustuLosc is closely related with A. palestinensis and that the former may be regarded as a junior synonym of the latter.

• Journal of Korea Multimedia Society
• /
• v.12 no.10
• /
• pp.1386-1396
• /
• 2009

Most decision support functions of spatial data warehouse rely on the OLAP operations upon a spatial cube. Meanwhile, higher performance is always guaranteed by indexing the cube, which stores huge amount of pre-aggregated information. Hierarchical Dwarf was proposed as a solution, which can be taken as an extension of the Dwarf, a compressed index for cube structures. However, it does not consider the spatial dimension and even aggregates incorrectly if there are redundant values at the lower levels. OLAP-favored Searching was proposed as a spatial hierarchy based OLAP operation, which employs the advantages of R-tree. Although it supports aggregating functions well against specified areas, it ignores the operations on the spatial dimensions. In this paper, an indexing approach, which aims at utilizing the concept hierarchy of the spatial cube for decision support, is proposed. The index consists of concept hierarchy trees of all dimensions, which are linked according to the tuples stored in the fact table. It saves storage cost by preventing identical trees from being created redundantly. Also, it reduces the OLAP operation cost by integrating the spatial and aspatial dimensions in the virtual concept hierarchy.

• Journal of the Korean earth science society
• /
• v.18 no.6
• /
• pp.553-558
• /
• 1997

Colorless and transparent cubic zirconia($Zr_{0.73}Y_{0.27}O_{1.87}$) crystal has been synthesized by the Bridgman-Stock-bager method(also called Skull melting method). $Y_2O_3$ is used as stabilizer. The crystal shows a vitreous luster with a slight oily appearance. Under a polarizing microscope, it shows isotropic nature with no appreciable anisotropism. Mohs hardness value is measured to be $8{\sim}8\frac{1}{2}$ and specific gravity 5.85. Under ultraviolet light it shows a faint white glow. The crystal structure of yttria stabilized zirconia was determined, using single crystal X-ray diffraction techniques to be a cubic symmetry, space group $Fm\overline{3}m({O^5}_h)$ with $a=5.1552(5){\AA}$, $V=136.99(5){\AA}^3$, Z=4, and R=0.0488 for 29 unique reflections. Each zirconium atom is at the center of eight oxygen atoms situated at the corners of a surrounding cube and each oxygen atom is at the center of a tetrahedron of zirconium atoms. So a coordination of 8:4 holds in the structure.