• Proceedings of the Korean Society of Computer Information Conference
• /
• 2011.01a
• /
• pp.85-88
• /
• 2011

Fuzzy C-Means (FCM) 알고리듬은 가장 널리 사용되는 군집화 알고리듬 중 하나로 다양한 응용 분야에서 사용되고 있다. 하지만 FCM은 여러 가지 문제점을 가지고 있으며 초기 원형 설정이 그 중 하나이다. FCM은 국부 최적해에 수렴하므로 초기 원형 설정에 따라 클러스터링 결과가 달라진다. 이 논문에서는 이러한 FCM의 초기 원형 설정 문제를 개선하기 위하여 커널밀도 추정 (kernel density estimation) 기법을 활용하는 방법을 제안한다. 제안한 방법에서는 먼저 커널 밀도 추정을 수행한 후 밀도가 높은 지역에 클러스터의 초기 원형을 설정하고 원형이 설정된 영역의 밀도를 감소시키는 과정을 반복함으로써 효율적으로 초기 원형을 설정할 수 있다. 제안된 방법이 일반적으로 사용되는 무작위 초기화 방법에 비해 효율적이라는 사실은 실험결과를 통해 확인할 수 있다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.8
• /
• pp.1659-1664
• /
• 2011

Fuzzy C-Means (FCM) is one of the most widely used clustering algorithms and has been used in many applications successfully. However, FCM has some shortcomings and initial prototype selection is one of them. As FCM is only guaranteed to converge on a local optimum, different initial prototype results in different clustering. Therefore, much care should be given to the selection of initial prototype. In this paper, a new initialization method for FCM using kernel density estimation (KDE) is proposed to resolve the initialization problem. KDE can be used to estimate non-parametric data distribution and is useful in estimating local density. After KDE, in the proposed method, one initial point is placed at the most dense region and the density of that region is reduced. By iterating the process, initial prototype can be obtained. The initial prototype such obtained showed better result than the randomly selected one commonly used in FCM, which was demonstrated by experimental results.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2014.05a
• /
• pp.76-78
• /
• 2014

Naive Bayes (NB, for shortly) learning is more popular, faster and effective supervised learning method to handle the labeled datasets especially in which have some noises, NB learning also has well performance. However, the conditional independent assumption of NB learning imposes some restriction on the property of handling data of real world. Some researchers proposed lots of methods to relax NB assumption, those methods also include attribute weighting, kernel density estimating. In this paper, we propose a novel approach called NB Based on Attribute Weighting in Kernel Density Estimation (NBAWKDE) to improve the NB learning classification ability via combining kernel density estimation and attribute weighting.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.1A
• /
• pp.35-40
• /
• 2011

The complex-valued blind algorithm based on a set of randomly generated symbols and Euclidean distance can take advantage of information theoretic learning and cope with the channel phase rotation problems. On the algorithm, in this paper, the effect of kernel size has been studied and a kernel-modified version of the algorithm that rearranges the forces between the information potentials by kernel-modification has been proposed. In simulation results for 16 QAM and complex-channel models, the proposed algorithm show significantly enhanced performance of symbol-point concentration and no phase rotation problems caused by the complex channel models.

• Journal for History of Mathematics
• /
• v.17 no.2
• /
• pp.15-20
• /
• 2004

We investigate the unbiasedness and consistency as the statistical properties of density estimators. We show histogram, kernel density estimation, and local adaptive smoothing as density smoothing in this paper. Also, the early and recent research on nonparametric density estimation is described and discussed.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.387-390
• /
• 2024

본 논문에서는 효율적인 2차 오차 함수를 이용하여 입자 기반에서 EMC(Extended Marching Cubes) 알고리즘을 구현할 수 있는 새로운 알고리즘을 제안한다. Smoothing 커널(Kernels)을 통해 계산한 입자 평균 위치에서 레벨셋(Level-set)을 계산해 스칼라장을 구축한다. 그리고 난 뒤 SPH(Smoothed particle hydrodynamics)기반의 커널을 통해 밀도, 입자 평균 위치를 계산한다. 스칼라장을 이용해 등가 곡면(Isosurface)을 찾고 음함수로 표현된 표면을 구성한다. SPH 커널을 공간에서 미분하면 공간상의 어느 위치에서나 기울기를 계산할 수 있고, 이를 통해 얻어진 법선벡터를 이용하여 일반적인 EMC나 DC(Dual contouring)에서 사용하는 2차 오차 함수를 효율적으로 설계한다. 결과적으로 제안하는 방법은 메쉬와 같이 연결정보다 없는 입자 기반 데이터에서도 EMC 알고리즘을 구현하여 볼륨(Volume) 손실을 줄이고, 복잡한 음함수 표면을 표현할 수 있게 한다.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2022.06a
• /
• pp.426-428
• /
• 2022

해상 교통량을 정량적으로 평가하고 추출하기 위한 방법으로 선박 AIS 데이터 기반의 밀집도 분석을 활용하고 있다. 밀집도는 단위시간 당 단위면적에 분포하는 선박 통항량을 계산한 것으로, 일반적으로 그리드 셀 내에 존재하는 선박 항적 포인트 개수, 항적도 라인 길이, 선박 척수 등을 계산한 밀집도 분석 방법과 커널 밀도 추정(Kernel Density Estimation) 방법 등이 있다. 하지만, AIS 데이터의 특징상 선박 속력에 따라 수신 주기가 다르기 때문에 항적이 등간격으로 나타나지 않는 문제점이 있으며, 선박의 이동과 시간의 속성으로 인해 각각의 밀집도 분석 방법은 한계점이 존재한다. 따라서 본 연구에서는 실측 AIS 데이터를 이용하여 다양한 방법의 선박 밀집도 분석을 수행하고 이를 비교하였다. 그 결과, 항적도 라인 길이에 의한 밀집도 분석이 가장 정량적인 방법으로 나타났으며 이를 통항 척수로 변환할 수 있는 선박 밀집도 분석을 개선방안으로 제안한다.

• Journal of IKEEE
• /
• v.13 no.4
• /
• pp.1-7
• /
• 2009

A video segmentation algorithm, which can extract objects even with non-stationary backgrounds, is proposed in this work. The proposed algorithm is composed of three steps. First, we perform an initial segmentation interactively to build the probability density functions of colors per each macro block via kernel density estimation. Then, for each subsequent frame, we construct a coherence strip, which is likely to contain the object contour, by exploiting spatio-temporal correlations. Finally, we perform the segmentation by minimizing an energy function composed of color, coherence, and smoothness terms. Experimental results on various test sequences show that the proposed algorithm provides accurate segmentation results.

• Journal of IKEEE
• /
• v.28 no.1
• /
• pp.6-13
• /
• 2024

This paper proposes an approach to detect abnormal data in automotive controller area network (CAN) using an unsupervised learning model, i.e. autoencoder and Gaussian kernel density estimation function. The proposed autoencoder model is trained with only message ID of CAN data frames. Afterwards, by employing the Gaussian kernel density estimation function, it effectively detects abnormal data based on the trained model characterized by the optimally determined number of frames and a loss threshold. It was verified and evaluated using four types of attack data, i.e. DoS attacks, gear spoofing attacks, RPM spoofing attacks, and fuzzy attacks. Compared with conventional unsupervised learning-based models, it has achieved over 99% detection performance across all evaluation metrics.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.4
• /
• pp.19-26
• /
• 2013

Pseudo sample neural network (PSNN) is a variant of traditional neural network using pseudo samples to mitigate the local-optima-convergence problem when the size of training samples is small. PSNN can take advantage of the smoothed solution space through the use of pseudo samples. PSNN has a focus on the quantity problem in training, whereas, methods stressing the quality of training samples is presented in this paper to improve further the performance of PSNN. It is evident that typical samples and highly correlated features help in training. In this paper, therefore, kernel density estimation is used to select typical samples and correlation factor is introduced to select features, which can improve the performance of PSNN. Debris flow data set is used to demonstrate the usefulness of the proposed methods.