• Title/Summary/Keyword: 데이터 샘플링

Search Result 510, Processing Time 0.028 seconds

Preprocessing in a Noninvasive Sensor System (비침습적 센서 시스템에서 전처리 연산)

  • Oh, Hyun-Kyo;Keum, Hyouseob;Cho, Seung-Ho;Kim, Heong-Tae;Moon, Bong-Hee
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2013.05a
    • /
    • pp.83-85
    • /
    • 2013
  • 본 연구는 사용자가 센서를 의식하지 않고 편안하게 일상생활을 영위할 수 있는 비침습적 방식의 센서를 활용하여 향후 침대 위에 있는 사람의 움직임을 정량적으로 측정하고자 한다. 이러한 목적으로 필름 형태의 압전센서를 사용하는 센서 시스템을 구축하였으며, 본 논문에서는 구축된 시스템에서 필요한 전처리 과정을 제시한다. 본 연구에서 사용된 압전센서는 침대 매트리스 아래에 설치하였다. 사람의 움직임에 의한 압전센서의 출력 신호를 증폭하고 샘플링하여 PC로 전송하는 컨트롤러, 컨트롤러로부터 센서 데이터를 수신하고, 사용자에게 센서 데이터를 시각적으로 제시하는 모니터링 프로그램을 개발하였다. 본 연구에서는 컨트롤러에서의 노이즈 제거, 증폭, 샘플링 등의 전처리, 모니터링 프로그램에 의해 수집된 센서 데이터에 대한 이동 평균 필터, 불필요한 움직임이 없는 구간을 제거 후 움직임이 있는 구간 추출 등의 전처리 과정을 제시한다. 이러한 전처리 연산은 향후 침대 위 인체의 움직임을 정량적으로 측정하고, 행동유형을 식별하는데 기여하게 될 것이다.

The Optimization of Ensembles for Bankruptcy Prediction (기업부도 예측 앙상블 모형의 최적화)

  • Myoung Jong Kim;Woo Seob Yun
    • Information Systems Review
    • /
    • v.24 no.1
    • /
    • pp.39-57
    • /
    • 2022
  • This paper proposes the GMOPTBoost algorithm to improve the performance of the AdaBoost algorithm for bankruptcy prediction in which class imbalance problem is inherent. AdaBoost algorithm has the advantage of providing a robust learning opportunity for misclassified samples. However, there is a limitation in addressing class imbalance problem because the concept of arithmetic mean accuracy is embedded in AdaBoost algorithm. GMOPTBoost can optimize the geometric mean accuracy and effectively solve the category imbalance problem by applying Gaussian gradient descent. The samples are constructed according to the following two phases. First, five class imbalance datasets are constructed to verify the effect of the class imbalance problem on the performance of the prediction model and the performance improvement effect of GMOPTBoost. Second, class balanced data are constituted through data sampling techniques to verify the performance improvement effect of GMOPTBoost. The main results of 30 times of cross-validation analyzes are as follows. First, the class imbalance problem degrades the performance of ensembles. Second, GMOPTBoost contributes to performance improvements of AdaBoost ensembles trained on imbalanced datasets. Third, Data sampling techniques have a positive impact on performance improvement. Finally, GMOPTBoost contributes to significant performance improvement of AdaBoost ensembles trained on balanced datasets.

Study on WAMAC System Architecture Design, Including PMU Data Verification System (PMU Data 검증시스템을 포함한 WAMAC 시스템 설계에 관한 연구)

  • Cho, Jun-Hee;Choi, Mi-Hwa;Lee, Myeong-Woo;Kim, Sang-Tae;Woo, Doug-Je
    • KIPS Transactions on Computer and Communication Systems
    • /
    • v.1 no.3
    • /
    • pp.181-186
    • /
    • 2012
  • PMU based power grid monitoring and control system, WAMAC (Wide Area Monitoring And Control) system is required system design for accurate power data without error and loss through a system-wide. In the paper, we propose system design that measured data from PMU transmitted without loss to PDC and DSM server. and we propose a method to verify the real-time "data has been transmitted accurately". Verification system has been designed to reflect the WAMAC system. Therefore the WAMAC can enhance the reliability of the analysis of the data, and it can monitor lossless real-time trend data.

An Efficient Angular Space Partitioning Based Skyline Query Processing Using Sampling-Based Pruning (데이터 샘플링 기반 프루닝 기법을 도입한 효율적인 각도 기반 공간 분할 병렬 스카이라인 질의 처리 기법)

  • Choi, Woosung;Kim, Minseok;Diana, Gromyko;Chung, Jaehwa;Jung, Soonyong
    • KIPS Transactions on Software and Data Engineering
    • /
    • v.6 no.1
    • /
    • pp.1-8
    • /
    • 2017
  • Given a multi-dimensional dataset of tuples, a skyline query returns a subset of tuples which are not 'dominated' by any other tuples. Skyline query is very useful in Big data analysis since it filters out uninteresting items. Much interest was devoted to the MapReduce-based parallel processing of skyline queries in large-scale distributed environment. There are three requirements to improve parallelism in MapReduced-based algorithms: (1) workload should be well balanced (2) avoid redundant computations (3) Optimize network communication cost. In this paper, we introduce MR-SEAP (MapReduce sample Skyline object Equality Angular Partitioning), an efficient angular space partitioning based skyline query processing using sampling-based pruning, which satisfies requirements above. We conduct an extensive experiment to evaluate MR-SEAP.

Location Estimation System based on Majority Sampling Data (머저리티 샘플링 데이터 기반 위치 추정시스템)

  • Park, Geon-Yeong;Jeon, Min-Ho;Oh, Chang-Heon
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.18 no.10
    • /
    • pp.2523-2529
    • /
    • 2014
  • Location estimation service can be provided outdoors using various location estimation system based on GPS. However, location estimation system is based on existing indoor resources as GPS cannot be used because of insufficient visible satellites and weak signals. The fingerprinting technique that uses WLAN signal, in particular, is good to use indoors because it uses RSSI provided by AP to estimate location. However, its accuracy may vary depending on how accurate data the offline stage used where the fingerprinting map is built. The study sampled various data at the stage that builds the fingerprinting map and suggested a location estimation system that enhances its precision by saving the data of high frequency among them to improve this problem. The suggested location estimation system based on majority sampling data estimates location by filtering RSSI data of the highest frequency at the client and server to be saved at a map, building the map and measuring a similar distance. As a result of the test, the location estimation precision stood at minimum 87.5 % and maximum 90.4% with the margin of error at minimum 0.25 to 2.72m.

Sample thread based real-time BRDF rendering (샘플 쓰레드 기반 실시간 BRDF 렌더링)

  • Kim, Soon-Hyun;Kyung, Min-Ho;Lee, Joo-Haeng
    • Journal of the Korea Computer Graphics Society
    • /
    • v.16 no.3
    • /
    • pp.1-10
    • /
    • 2010
  • In this paper, we propose a novel noiseless method of BRDF rendering on a GPU in real-time. Illumination at a surface point is formulated as an integral of BRDF producted with incident radiance over the hemi-sphere domain. The most popular method to compute the integral is the Monte Carlo method, which needs a large number of samples to achieve good image quality. But, it leads to increase of rendering time. Otherwise, a small number of sample points cause serious image noise. The main contribution of our work is a new importance sampling scheme producing a set of incoming ray samples varying continuously with respect to the eye ray. An incoming ray is importance-based sampled at different latitude angles of the eye ray, and then the ray samples are linearly connected to form a curve, called a thread. These threads give continuously moving incident rays for eye ray change, so they do not make image noise. Since even a small number of threads can achieve a plausible quality and also can be precomputed before rendering, they enable real-time BRDF rendering on the GPU.

Sound Enhancement of low Sample rate Audio Using LMS in DWT Domain (DWT영역에서 LMS를 이용한 저 샘플링 비율 오디오 신호의 음질 향상)

  • 백수진;윤원중;박규식
    • The Journal of the Acoustical Society of Korea
    • /
    • v.23 no.1
    • /
    • pp.54-60
    • /
    • 2004
  • In order to mitigate the problems in storage space and network bandwidth for the full CD quality audio, current digital audio is always restricted by sampling rate and bandwidth. This restriction normally results in low sample rate audio or calls for the data compression scheme such as MP3. However, they can only reproduce a lower frequency range than a regular CD quality because of the Nyquist sampling theory. Consequently they lose rich spatial information embedded in high frequency. The propose of this paper is to propose efficient high frequency enhancement of low sample rate audio using n adaptive filtering and DWT analysis and synthesis. The proposed algorithm uses the LMS adaptive algorithm to estimate the missing high frequency contents in DWT domain and it then reconstructs the spectrally enhanced audio by using the DWT synthesis procedure. Several experiments with real speech and audio are performed and compared with other algorithm. From the experimental results of spectrogram and sonic test, we confirm that the proposed algorithm outperforms the other algorithm and reasonably works well for the most of audio cases.

Performance Improvement of Fractal Dimension Estimator Based on a New Sampling Method (새로운 샘플링법에 기초한 프랙탈 차원 추정자의 정도 개선)

  • Jin, Gang-Gyoo;Choi, Dong-Sik
    • Journal of Navigation and Port Research
    • /
    • v.38 no.1
    • /
    • pp.45-52
    • /
    • 2014
  • Fractal theory has been widely used to quantify the complexity of remotely sensed digital elevation models and images. Despite successful applications of fractals to a variety of fields including computer graphics, engineering and geosciences, the performance of fractal estimators depends highly on data sampling. In this paper, we propose an algorithm for computing the fractal dimension based on the triangular prism method and a new sampling method. The proposed sampling method combines existing two methods, that is, the geometric step method and the divisor step method to increase pixel utilization. In addition, while the existing estimation methods are based on $N{\times}M$ window, the proposed method expands to $N{\times}M$ window. The proposed method is applied to generated fractal DEM, Brodatz's image DB and real images taken in the campus to demonstrate its feasibility.

Sampling Techniques for Wireless Data Broadcast in Communication (통신에서의 무선 데이터 방송을 위한 샘플링 기법)

  • Lee, Sun Yui;Park, Gooman;Kim, Jin Young
    • Journal of Satellite, Information and Communications
    • /
    • v.10 no.3
    • /
    • pp.57-61
    • /
    • 2015
  • This paper describes the basic principles of 3D broadcast system and proposes new 3D broadcast technology that reduces the amount of data by applying CS(Compressed Sensing). Differences between Sampling theory and the CS technology concept was described. CS algorithm SS-CoSaMP(Single-Space Compressive Sampling Matched Pursuit) and AMP(Approximate Message Passing) was described. Image data compressed and restored by these algorithm was compared. Calculation time of the algorithm having a low complexity is determined.

A Study on the Scene Chang Detection Retrieval of Video Using MSE (MSE를 이용한 비디오의 장면전환 검색에 관한 연구)

  • 김단환;김형균;고석만;오무송
    • Proceedings of the Korea Multimedia Society Conference
    • /
    • 2002.05d
    • /
    • pp.1052-1055
    • /
    • 2002
  • 동영상 정보는 영상정보뿐만이 아니라 음성정보, 문자정보 및 각종 의미있는 정보들을 포함하고 있어서 기존의 검색 방법으로는 사용자가 원하는 이미지를 찾는데 어려움이 따른다. 따라서, 본 연구에서는 동영상 정보의 효율적인 활용을 위한 색인 방법으로 MSE(Mean Square Error) 도입하여 동영상의 장면전환 검색하는 방법을 제안한다. 이것은 영상 데이터를 대각선 방향으로 일정픽셀의 칼라 값을 추출하여 행렬A에 i×j행렬로 i는 프레임 수, j는 프레임의 영상 높이로 저장하고 동영상의 전체 구조를 파악할 수 있도록 정지영상으로 샘플링 하였다. 샘플링 된 데이터는 대용량 동영상 데이터 이용에 있어서 사용자가 전체 동영상의 장면전환점을 한눈으로 파악할 수 있고, 각 프레임의 MSE와 임계값을 초과하면 그 프레임이 장면전환점으로 검색한다.

  • PDF