• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2006년도 Proceedings of ISRS 2006 PORSEC Volume II
• /
• pp.635-638
• /
• 2006

Regression of conventional prediction techniques in data mining uses the model which is generated from the training step. This model is applied to new input data without any change. If this model is applied directly to time series, the rate of prediction accuracy will be decreased. This paper proposes an incremental regression for time series prediction like typhoon track prediction. This technique considers the characteristic of time series which may be changed over time. It is composed of two steps. The first step executes a fractional process for applying input data to the regression model. The second step updates the model by using its information as new data. Additionally, the model is maintained by only recent data in a queue. This approach has the following two advantages. It maintains the minimum information of the model by using a matrix, so space complexity is reduced. Moreover, it prevents the increment of error rate by updating the model over time. Accuracy rate of the proposed method is measured by RME(Relative Mean Error) and RMSE(Root Mean Square Error). The results of typhoon track prediction experiment are performed by the proposed technique IMLR(Incremental Multiple Linear Regression) is more efficient than those of MLR(Multiple Linear Regression) and SVR(Support Vector Regression).

• Journal of Information Processing Systems
• /
• 제14권4호
• /
• pp.851-876
• /
• 2018

Dynamic time warping (DTW) is the main algorithms for time series alignment. However, it is unsuitable for quasi-periodic time series. In the current situation, except the recently published the shape exchange algorithm (SEA) method and its derivatives, no other technique is able to handle alignment of this type of very complex time series. In this work, we propose a novel algorithm that combines the advantages of the SEA and the DTW methods. Our main contribution consists in the elevation of the DTW power of alignment from the lowest level (Class A, non-periodic time series) to the highest level (Class C, multiple-periods time series containing different number of periods each), according to the recent classification of time series alignment methods proposed by Boucheham (Int J Mach Learn Cybern, vol. 4, no. 5, pp. 537-550, 2013). The new method (quasi-periodic dynamic time warping [QP-DTW]) was compared to both SEA and DTW methods on electrocardiogram (ECG) time series, selected from the Massachusetts Institute of Technology - Beth Israel Hospital (MIT-BIH) public database and from the PTB Diagnostic ECG Database. Results show that the proposed algorithm is more effective than DTW and SEA in terms of alignment accuracy on both qualitative and quantitative levels. Therefore, QP-DTW would potentially be more suitable for many applications related to time series (e.g., data mining, pattern recognition, search/retrieval, motif discovery, classification, etc.).

• Wind and Structures
• /
• 제6권2호
• /
• pp.107-126
• /
• 2003

This paper describes a simple and practical approach through the application of Linear Stochastic Estimation (LSE) to reconstruct wind-induced pressure time series from the covariance matrix for structural load analyses on a low building roof. The main application of this work would be the reduction of the data storage requirements for the NIST aerodynamic database. The approach is based on the assumption that a random pressure field can be estimated as a linear combination of some other known pressure time series by truncating nonlinear terms of a Taylor series expansion. Covariances between pressure time series to be simulated and reference time series are used to calculate the estimation coefficients. The performance using different LSE schemes with selected reference time series is demonstrated by the reconstruction of structural load time series in a corner bay for three typical wind directions. It is shown that LSE can simulate structural load time series accurately, given a handful of reference pressure taps (or even a single tap). The performance of LSE depends on the choice of the reference time series, which should be determined by considering the balance between the accuracy, data-storage requirements and the complexity of the approach. The approach should only be used for the determination of structural loads, since individual reconstructed pressure time series (for local load analyses) will have larger errors associated with them.

• 한국정보과학회논문지:데이타베이스
• /
• 제33권2호
• /
• pp.201-213
• /
• 2006

일정 기간 동안 객체의 변화한 값들을 기록한 것을 그 객체에 대한 시계열 데이타 시퀀스라고 부르며, 이들의 집합을 시계열 데이타베이스라고 한다. 서브시퀀스 매칭은 주어진 질의 시퀀스와 변화의 추세가 유사한 서브시퀀스들을 시계열 데이타베이스로부터 검색하는 연산이다. 본 논문에서는 서브시퀀스 매칭의 성능을 극대화하기 위한 방안을 제시한다. 먼저, 윈도우 크기 효과로 인한 서브시퀀스 매칭의 심각한 성능 저하 현상을 정량적으로 관찰하여, 하나의 윈도우 크기를 대상으로 만든 단 하나의 인덱스만을 이용하는 것은 실제 응용에서 만족할만한 성능을 제공할 수 없다는 것을 규명하였다 또한, 이러한 문제로 인해 다양한 윈도우 크기들을 기반으로 다수의 인덱스들을 구성하여 서브시퀀스 매칭을 수행하는 인덱스 보간법의 응용이 필요함을 보였다. 인덱스 보간법을 응용하여 서브시퀀스 매칭을 수행하기 위해서는 먼저 다수의 인덱스들을 위한 윈도우 크기들을 결정해야 한다. 본 연구에서는 물리적 데이타베이스 설계 방식을 이용하여 이러한 최적의 다수의 윈도우 크기들을 선정하는 문제를 해결하였다. 이를 위하여 시계열 데이터 베이스에서 수행될 예정인 질의 시퀀스들의 집합과 인덱스 구성의 기반이 되는 윈도우들의 크기의 집합이 주어질 때, 전체 서브시퀀스 매칭들을 수행하는 데에 소요되는 비용을 예측할 수 있는 공식을 산출하였다. 또한, 이 비용 공식을 이용하여 전체 서브시퀀스 매칭들의 성능을 극대화 할 수 있는 최적의 윈도우 크기들을 결정하는 알고리즘을 제안하였으며, 이 알고리즘의 최적성과 효율성을 이론적으로 규명하였다. 끝으로, 실제 주식 데이타와 대량의 합성 데이타를 이용한 실험 결과, 제안된 기법은 기존의 단순한 기법과 비교하여 1.5배에서 7.8배 성능이 향상됨을 보였다.

• 천문학회보
• /
• 제42권2호
• /
• pp.88.3-89
• /
• 2017

We present our experiences with open-source NoSQL database systems in analyzing spatial and temporal astronomical data. We conduct experiments of using Redis in-memory NoSQL database system by modifying and exploiting its support of geohash for astronmical spatial data. Our experiment focuses on performance, cost, difficulty, and scalability of the database system. We also test OpenTSDB as a possible NoSQL database system to process astronomical time-series data. Our experiments include ingesting, indexing, and querying millions or billions of astronomical time-series measurements. We choose our KMTNet data and the public VVV (VISTA Variables in the Via Lactea) catalogs as test data. We discuss issues in using these NoSQL database systems in astronomy.

• 한국콘텐츠학회:학술대회논문집
• /
• 한국콘텐츠학회 2003년도 추계종합학술대회 논문집
• /
• pp.290-297
• /
• 2003

시계열 데이터베이스란 객체의 변화되는 값들의 연속으로 구성된 데이터 시퀀스들의 집합이며, 타임 워핑 하의 서브시퀀스 매칭은 주어진 질의 시퀀스와 타임 워핑 거리가 허용치 이하인 서브시퀀스들을 시계열 데이터베이스로부터 찾아내는 연산이다. 본 논문에서는 먼저 타임 워핑 하의 시퀀스 매칭을 지원하는 기존의 기법들의 특성을 지적하고, 이들을 전체매칭 및 서브시퀀스 매칭에 각각 적용하는 방안에 관하여 논의한다. 또한, 실제 주식 데이터를 이용한 다양한 실험을 통하여 이들에 대한 정량적인 성능평가를 수행한다. 타임 워핑 하의 서브시퀀스 매칭을 위한 기존 기법들의 성능을 상호 비교한 연구 결과는 아직 제시된 바 없다. 따라서 본 연구 결과는 이러한 세 가지 기법들에 대한 성능을 제시하는 좋은 자료로서 사용될 수 있을 것이다.

• 한국전자거래학회:학술대회논문집
• /
• 한국전자거래학회 2005년도 종합학술대회
• /
• pp.147-153
• /
• 2005

최근 인터넷 쇼핑몰에서 상품을 구매하는 고객들에게 편의성과 효율성을 제공하기 위하여 구매자들의 선호도나 가격에 맞는 상품을 추천해 주는 연구들이 활발하게 진행되고 있지만추천된 상품들의 구매시점에 관한 연구는 찾아보기 어렵다. 이에 본 논문에서는 인터넷 쇼핑몰의 적극적인 마케팅 일환으로 판매가격의 흐름을 시계열 패턴으로 분석하여 상품의 구매시점 정보를 제공하는 방안을 제안한다. 이를 위하여 과거의 판매 기록 데이터베이스에 있는 판매가격의 기준이 되는 패턴과 유사한 변화를 보이는 패턴을 정규화된 유사도로써 검색하고, 검색된 가격 패턴을 기준으로 미래의 가격 패턴의 변화를 분석하여, 미래 가격 패턴의 변화 폭에 따라 상품에 대한 구매시점을 제공한다.

• 반도체디스플레이기술학회지
• /
• 제18권4호
• /
• pp.57-61
• /
• 2019

Nowadays, portable digital cameras such as smart phone cameras are being popularly used for entertainment and visual information recording. Given a database of geo-tagged images, a visual location recognition system can determine the place depicted in a query photo. One of the most common visual location recognition approaches is the bag-of-words method where local image features are clustered into visual words. In this paper, we propose a new bag-of-words-based visual location recognition algorithm using time-series streetview database. The proposed algorithm selects only a small subset of image features which will be used in image retrieval process. By reducing the number of features to be used, the proposed algorithm can reduce the memory requirement of the image database and accelerate the retrieval process.

• 한국경영과학회지
• /
• 제21권3호
• /
• pp.197-214
• /
• 1996

Financial analysis of stock data usually involves extensive computation of large amount of time series data sets. To handle the large size of the data sets and complexity of the analyses, database management systems have been increasingly adaopted for efficient management of stock data. Specially, relational database management system is employed more widely due to its simplistic data management approach. However, the normalized two-dimensional tables and the structured query language of the relational system turn out to be less effective than expected in accommodating time series stock data as well as the various computational operations. This paper explores a new data management approach to stock data management on the basis of an object-oriented database management system (ODBMS), and proposes a data model supporting times series data storage and incorporating a set of financial analysis functions. In terms of functional stock data analysis, it specially focuses on a primitive set of operations such as variance of stock data. In accomplishing this, we first point out the problems of a relational approach to the management of stock data and show the strength of the ODBMS. We secondly propose an object model delineating the structural relationships among objects used in the stock data management and behavioral operations involved in the financial analysis. A prototype system is developed using a commercial ODBMS.

• 한국컴퓨터정보학회논문지
• /
• 제24권5호
• /
• pp.35-40
• /
• 2019

In this paper, we design and implement a partial denoising boundary matching system using indexing techniques. Converting boundary images to time-series makes it feasible to perform a fast search using indexes even on a very large image database. Thus, using this converting method we develop a client-server system based on the previous partial denoising research in the GUI(graphical user interface) environment. The client first converts a query image given by a user to a time-series and sends denoising parameters and the tolerance with this time-series to the server. The server identifies similar images from the index by evaluating a range query, which is constructed using inputs given from the client and sends the resulting images to the client. Experimental results show that our system provides many intuitive and accurate matching results.