• Journal of KIISE
• /
• v.42 no.1
• /
• pp.86-92
• /
• 2015

A level set tree provides a useful representation of a multidimensional density function. Visualizing the data structure as a tree offers many advantages for data analysis and clustering. In this paper, we present a level set tree estimation algorithm for use with a set of data points. The proposed algorithm creates a level set tree from a family of level sets estimated over a whole range of levels from zero to infinity. Instead of estimating density function then thresholding, we directly estimate the density level sets using one-class support vector machines (OC-SVMs). The level set estimation is facilitated by the OC-SVM solution path algorithm. We demonstrate the proposed level set tree algorithm on benchmark data sets.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.6
• /
• pp.208-221
• /
• 2020

The objective of this study is to estimate and analyze the traffic density of continuous flow using the trajectory of individual vehicles and the headway of sample probe vehicles-front vehicles obtained from ADAS (Advanced Driver Assitance System) installed in sample probe vehicles. In the past, traffic density of continuous traffic flow was mainly estimated by processing data such as traffic volume, speed, and share collected from Vehicle Detection System, or by counting the number of vehicles directly using video information such as CCTV. This method showed the limitation of spatial limitations in estimating traffic density, and low reliability of estimation in the event of traffic congestion. To overcome the limitations of prior research, In this study, individual vehicle trajectory data and vehicle headway information collected from ADAS are used to detect the space on the road and to estimate the spatiotemporal traffic density using the Generalized Density formula. As a result, an analysis of the accuracy of the traffic density estimates according to the sampling rate of ADAS vehicles showed that the expected sampling rate of 30% was approximately 90% consistent with the actual traffic density. This study contribute to efficient traffic operation management by estimating reliable traffic density in road situations where ADAS and autonomous vehicles are mixed.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2018.11a
• /
• pp.275-277
• /
• 2018

선박 및 해상교통관제에 있어서 교통 혼잡구역에 대한 선박교통밀도 예측은 선박충돌사고 예방에 중요하다. 선박 교통밀도 예측정보는 사전에 진입하는 선박들에게 속력조정, 우회항로 이용 등 사전 조치가 가능하다. 본 연구에서는 해상 선박교통상황을 딥러닝 네트워크에 학습한 주의구역 선박교통류 예측 모델(Ship Traffic Extraction Network, STENet)을 제안하여 주의구역의 선박교통류 예측을 수행하고자 한다. STENet 모델 학습을 위해 여수해역 AIS 데이터를 전처리하고, 생성된 입력(해상교통상황)-출력(주의구역 교통밀도) 쌍 데이터를 적용하여 STENet 모델을 학습하였다. 학습된 모델을 이용하여 선박교통류 예측을 한 결과, 중기예측은 표준 절대 오차(mean absolute error)가 0.4-0.5척이 였으며, 장기예측은 0.7-0.8척의 오차로 기존의 Dead Reckoning에 의한 방법보다 50% 이상 교통밀도 예측성능이 향상 되었다.

• Journal of KIISE:Databases
• /
• v.36 no.4
• /
• pp.247-253
• /
• 2009

Rank learning and processing have gained much attention in the IR and data mining communities for the last decade. While other data mining techniques such as classification and regression have been actively researched to interoperate with RDBMS by using the tightly coupled or loose coupling approaches, ranking has been researched independently without integrating into RDBMS. This paper proposes a tightly coupled integration of the Ranking SVM into MySQL in order to perform the rank learning task efficiently within the RDBMS. We implemented new SQL commands for learning ranking functions and predicting ranking scores. We evaluated our tightly coupled integration of Ranking SVM by comparing it to a loose coupling implementation. The experiment results show that our approach has a performance improvement of $10{\sim}40%$ in the training phase and 60% in the prediction phase.

• Broadcasting and Media Magazine
• /
• v.25 no.3
• /
• pp.38-48
• /
• 2020

국내에서는 차세대 방송을 위해 ATSC 3.0 기술을 기반으로 한 UHD 방송이 상용화되었다. ATSC 3.0은 IP 프로토콜을 중심으로 통신망과 쉽게 융합 가능하도록 고안된 방송 기술로 고품질 UHD 방송을 포함하여 DMB와 같은 모바일 TV 및 통신망과 연동할 경우 다양한 부가데이터 서비스 제공이 가능한 기술이다. 대표적인 부가 데이터 서비스로는 App 서비스가 있지만 비디오와 무관한 데이터 서비스로 차량 운전자를 위한 여행 및 교통 정보 즉 TTI 서비스가 있다. TTI 서비스는 차량을 목표로 하는 여행 및 교통 정보를 방송망과 더불어 통신망을 통해 전송하여 일반 차량을 포함하여 인터넷이 가능한 커넥티드카 및 자율주행차량에 요구되는 고정밀 위치정보, C-ITS 정보 및 대용량 고정밀 지도 데이터 등을 전달할 수 있다. 본 고에서는 ATSC 3.0 기반의 UHD 방송망과 브로드밴드 망을 활용하여 차량에 유익한 정보를 제공할 수 있는 TTI 서비스에 대해 소개하고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2005.11a
• /
• pp.101-104
• /
• 2005

최근 GIS 시스템, 위성사진, 원격 탐사 시스템과 같은 다양한 응용 시스템으로부터 수집된 방대한 양의 공간 데이터에서 지식을 발견하는 공간 데이터 마이닝에 대한 관심이 더욱 높아지고 있다. 기존의 공간 데이터마이닝에 대한 연구들은 방대한 비공간 데이터들의 지식을 효율적으로 탐사하고자 하였다. 그러나 기존의 시스템은 발견된 지식의 효과성을 보장하지 못하는 문제점을 가진다. 따라서 본 논문은 공간 데이터 타입을 포함하는 대용량의 데이터들로부터 효과성을 보장하는 특성화 지식 탐사시스템을 제안한다. 본 논문에서 제안하는 공간 특성화 지식 탐사시스템은 밀도 기반의 클러스터링 기법을 적용하여 탐사된 특성화 지식의 효과성을 높였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11b
• /
• pp.148-150
• /
• 2005

최근 유비쿼터스 컴퓨팅의 관심이 증대되면서, 방대하고 다양한 형태의 데이터에 대한 효율성과 효과성을 고려한 지식 탐사방법연구의 필요성이 제기되었다. 기존의 지식 탐사방법에 대한 연구들은 방대한 비공간 데이터들의 지식을 효율적으로 탐사하고자 하였다. 그러나 기존의 연구는 탐사된 지식의 효율성안을 고려하여 유용한 지식탐사를 보장하지 못하는 문제점을 가진다. 따라서 본 논문은 공간 데이터 타입을 포함하는 대용랑의 데이터들로부터 효과성을 보장하는 특성화 지식 탐사방법을 제안한다. 본 논문에서 제안하는 특성화 지식 탐사기법은 공간 및 비공간 데이터들의 특성을 나타내는 요약된 지식을 제공하며, 밀도 기반의 클러스터링 기법을 적용하여 특성화 지식 탐사의 효과성을 높인다.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.43 no.11 s.353
• /
• pp.98-104
• /
• 2006

The conventional burst-mode APC(Automatic Power Control) circuit had an effective structure that was suitable for a low power consumption and a monolithic chip. However, as data rate was increased, it caused errors due to the effect of the zero density. In this paper, we invented a new structured peak-comparator which could compensate the unbalance of the injected currents using double gated MOS and MOS diode. And we proposed a new burst-mode APC adopting it. The new peak-comparator in the proposed APC was very robust to zero density variations maintaining the correct decision point of the current comparison at high data rate. It was also suitable for a low power consumption and a monolithic chip due to lack of large capacitors.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.21 no.4
• /
• pp.309-316
• /
• 2003

Rapid and accurate data processing is required in many GPS(Global Positioning System) applications including surveying. While one can use four different kinds of GPS satellite orbits, we evaluated the accuracy and precision of each kind of orbits to find the best candidate for rapid and accurate data processing. The four different kinds of orbits we: broadcast orbits from GPS satellites; and ultra-rapid orbits, rapid orbits, and precise orbits provided by international GPS data analysis centers such as IGS. With GIPSY and ultra-rapid orbits, we could get the positioning accuracy of 1.5cm from seven days of GPS data. From this study, we conclude that rapid and accurate data processing is achieved with GIPSY and ultra-rapid orbits.

• The KIPS Transactions:PartD
• /
• v.13D no.5 s.108
• /
• pp.633-640
• /
• 2006

Spatial clustering, which groups similar objects based on their distance, connectivity, or their relative density in space, is an important component of spatial data mining. Density-based and grid-based clustering are two main clustering approaches. The former is famous for its capability of discovering clusters of various shapes and eliminating noises, while the latter is well known for its high speed. Clustering large data sets has always been a serious challenge for clustering algorithms, because huge data set would make the clustering process extremely costly. In this paper, we propose an enhanced Density-Grid based Clustering algorithm for Large spatial database by setting a default number of intervals and removing the outliers effectively with the help of a proper measurement to identify areas of high density in the input data space. We use a density threshold DT to recognize dense cells before neighbor dense cells are combined to form clusters. When proposed algorithm is performed on large dataset, a proper granularity of each dimension in data space and a density threshold for recognizing dense areas can improve the performance of this algorithm. We combine grid-based and density-based methods together to not only increase the efficiency but also find clusters with arbitrary shape. Synthetic datasets are used for experimental evaluation which shows that proposed method has high performance and accuracy in the experiments.