• Journal of IKEEE
• /
• v.15 no.3
• /
• pp.191-197
• /
• 2011

An image segmentation model using fuzzy c-means with intuition (FCM-I) model is proposed for the segmentation of magnetic resonance image in this paper. In FCM-I, a measurement called intuition level is adopted so that the intuition level helps to alleviate the effect of noises. A practical magnetic resonance image data set is used for image segmentation experiment and the performance is compared with those of some conventional algorithms. Results show that the segmentation method based on FCM-I compares favorably to several conventional clustering algorithms. Since FCM-I produces cluster prototypes less sensitive to noises and to the selection of involved parameters than the other algorithms, FCM-I is a good candidate for image segmentation problems.

• Proceedings of the Korean Environmental Sciences Society Conference
• /
• 2003.11a
• /
• pp.33-36
• /
• 2003

Principal component analysis (PCA) was used to improve image classification by the unsupervised classification techniques, the K-means. To do this, I selected a Landsat TM scene of Jeju Island, Korea and proposed two methods for PCA： unstandardized PCA (UPCA) and standardized PCA (SPCA). The estimated accuracy of the image classification of Jeju area was computed by error matrix. The error matrix was derived from three unsupervised classification methods. Error matrices indicated that classifications done on the first three principal components for UPCA and SPCA of the scene were more accurate than those done on the seven bands of TM data and that also the results of UPCA and SPCA were better than those of the raw Landsat TM data. The classification of TM data by the K-means algorithm was particularly poor at distinguishing different land covers on the island. From the classification results, we also found that the principal component based classifications had characteristics independent of the unsupervised techniques (numerical algorithms) while the TM data based classifications were very dependent upon the techniques. This means that PCA data has uniform characteristics for image classification that are less affected by choice of classification scheme. In the results, we also found that UPCA results are better than SPCA since UPCA has wider range of digital number of an image.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.2
• /
• pp.57-64
• /
• 2014

Most previous studies of software fault prediction model which determines the fault-proneness of input modules have focused on supervised learning model using training data set. However, Unsupervised learning model is needed in case supervised learning model cannot be applied: either past training data set is not present or even though there exists data set, current project type is changed. Building an unsupervised learning model is extremely difficult that is why only a few studies exist. In this paper, we build unsupervised models using representative clustering algorithms, EM and DBSCAN, that have not been used in prior studies and compare these models with the previous model using K-means algorithm. The results of our study show that the EM model performs slightly better than the K-means model in terms of error rate and these two models significantly outperform the DBSCAN model.

• Earthquakes and Structures
• /
• v.12 no.5
• /
• pp.479-487
• /
• 2017

Frictional dampers are used in structural engineering as means of passive control. Meanwhile, frictional damper shave a disadvantage compared to viscous rivals since the slippage force must be exceeded to activate the device, and cannot be ideal full range of possible events. The concept of semi-active control is utilized to overcome this shortcoming. In this paper, a new semi-active frictional damper called Smart Adjustable Frictional (SAF) damper is introduced. SAF damper consists of hydraulic, electronic units and sensors which are all linked with an active control discipline. SAF acts as a smart damper which can adapt its slippage threshold during a dynamic excitation by measuring and controlling the structural response. The novelty of this damper is, while it controls the response of the structure in real time with acceptable time delay. The paper also reports on the results of a series of experiments which have been performed on SAF dampers to obtain their prescribed hysteretic behavior for various control algorithms. The results show that SAF can produce the desired slippage load of various algorithms in real time. Numerical models incorporating control simulations are also made to obtain the hysteretic response of the system which agrees closely with test results.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.2
• /
• pp.416-424
• /
• 2017

In this study, we propose a design of digits recognition system based on RBFNNs through a comparative study of pre-processing algorithms in order to recognize digits in handwritten. Histogram of Oriented Gradient(HOG) is used to get the features of digits in the proposed digits recognition system. In the pre-processing part, a dimensional reduction is executed by using Principal Component Analysis(PCA) and (2D)2PCA which are widely adopted methods in order to minimize a loss of the information during the reduction process of feature space. Also, The architecture of radial basis function neural networks consists of three functional modules such as condition, conclusion, and inference part. In the condition part, the input space is partitioned with the use of fuzzy clustering realized by means of the Fuzzy C-Means algorithm. Also, it is used instead of gaussian function to consider the characteristic of input data. In the conclusion part, the connection weights are used as the extended type of polynomial expression such as constant, linear, quadratic and modified quadratic. By using MNIST handwritten digit benchmarking database, experimental results show the effectiveness and efficiency of proposed digit recognition system when compared with other studies.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2006.11a
• /
• pp.421-426
• /
• 2006

We propose a new algorithm for K-medoids clustering which runs like the K-means clustering algorithm and test several methods for selecting initial medoids. The proposed algorithm calculates similarity matrix once and uses it for finding new medoids at every iterative step. To evaluate the proposed algorithm we use real and artificial data and compare with the clustering results of other algorithms in terms of three performance measures. Experimental results show that the proposed algorithm takes the reduced time in computation with comparable performance as compared to the Partitioning Around Medoids.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.10
• /
• pp.55-62
• /
• 2016

In this paper, we present indoor localization technologies using the bluetooth signal categorizing them into proximity based, triangulation based and fingerprinting based technologies. Then we provide localization accuracy improvement algorithms such as moving average, K-means, particle filter, and K-Nearest neighbor algorithms. We define important performance issues for indoor localization technologies and analyze recent technologies according to the performance issues. Finally we provide experimental results for correlation between RSSI and distance. We believe that this paper provide wise view and necessary information for recent localization technologies using the bluetooth signal.

• International Journal of Reliability and Applications
• /
• v.10 no.1
• /
• pp.1-15
• /
• 2009

Even though the impact of manufacturing quality to reliability is not considered much as well as that of design area, a major cause of an early failure of the product is known as manufacturing problem. This research applies two different types of neural network algorithms, the Back propagation (BP) algorithm and Learning Vector Quantization (LVQ) algorithm, to identify and classify the nonrandom variation pattern on the control chart based on knowledge-based diagnosis of dimensional variation. The performance and efficiency of both algorithms are evaluated to choose the better pattern recognition system for auto body assembly process. To analyze hundred percent of the data obtained by Optical Coordinate Measurement Machine (OCMM), this research considers an application in which individual observations rather than subsample means are used. A case study for analysis of OCMM data in underbody assembly process is presented to demonstrate the proposed knowledge-based pattern recognition system.

• Asian-Australasian Journal of Animal Sciences
• /
• v.15 no.7
• /
• pp.925-931
• /
• 2002

Gibbs sampling algorithms were implemented to the multi-trait threshold animal models with any combinations of multiple binary, ordered categorical, and linear traits and investigate the amount of bias on these models with two kinds of parameterization and algorithms for generating underlying liabilities. Statistical models which included additive genetic and residual effects as random and contemporary group effects as fixed were considered on the models using simulated data. The fully conditional posterior means of heritabilities and genetic (residual) correlations were calculated from 1,000 samples retained every 10th samples after 15,000 samples discarded as "burn-in" period. Under the models considered, several combinations of three traits with binary, multiple ordered categories, and continuous were analyzed. Five replicates were carried out. Estimates for heritabilities and genetic (residual) correlations as the posterior means were unbiased when underlying liabilities for a categorical trait were generated given by underlying liabilities of the other traits and threshold estimates were rescaled. Otherwise, when parameterizing threshold of zero and residual variance of one for binary traits, heritability estimates were inflated 7-10% upward. Genetic correlation estimates were biased upward if positively correlated and downward if negatively correlated when underling liabilities were generated without accounting for correlated traits on prior information. Residual correlation estimates were, consequently, much biased downward if positively correlated and upward if negatively correlated in that case. The more categorical trait had categories, the better mixing rate was shown.

• KIEE International Transactions on Power Engineering
• /
• v.5A no.2
• /
• pp.138-143
• /
• 2005

During faults, the voltage and current signals available to the relay are affected by the decaying dc component and harmonics. In order to make appropriate and accurate decisions, most of the relaying algorithms require the fundamental frequency phasor information that is immune to decaying dc effect and harmonics. The conventional Fourier ph as or estimation algorithm is affected by the presence of decaying-exponential transients in the fault signal. This paper presents a modified Fourier algorithm, which effectively eliminates the decaying dc component and the harmonics present in the fault signal. The decaying dc parameters are estimated by means of an out-of-band filtering technique. The decaying dc offset and harmonics are removed by means of a simple computational procedure that involves the design of two sets of Orthogonal digital OFT filters tuned at different frequencies and by creating three off-line look-up tables. The technique was tested for different decay rates of the decaying dc component. It was also compared with the conventional mimic plus the full cycle OFT algorithm. The results indicate that the proposed technique has a faster convergence to the desired value compared to the conventional mimic plus OFT algorithms over a wide range of decay rates. In all cases, the convergence to the desired value was achieved within one cycle of the power system frequency.