• Communications for Statistical Applications and Methods
• /
• v.28 no.1
• /
• pp.89-97
• /
• 2021

Due to boundedness and sum constraint, compositional data are often transformed by logratio transformation and their transformed data are put into traditional binary classification or discriminant analysis. However, it may be problematic to directly apply traditional multivariate approaches to the transformed data because class distributions are not Gaussian and Bayes decision boundary are not polynomial on the transformed space. In this study, we propose to use flexible classification approaches to transformed data for compositional data classification. Empirical studies using synthetic and real examples demonstrate that flexible approaches outperform traditional multivariate classification or discriminant analysis.

• International journal of advanced smart convergence
• /
• v.8 no.4
• /
• pp.75-81
• /
• 2019

Though recent advancement of deep learning methods have provided satisfactory results from large data domain, somehow yield poor performance on few-shot classification tasks. In order to train a model with strong performance, i.e. deep convolutional neural network, it depends heavily on huge dataset and the labeled classes of the dataset can be extremely humongous. The cost of human annotation and scarcity of the data among the classes have drastically limited the capability of current image classification model. On the contrary, humans are excellent in terms of learning or recognizing new unseen classes with merely small set of labeled examples. Few-shot learning aims to train a classification model with limited labeled samples to recognize new classes that have neverseen during training process. In this paper, we increase the backbone depth of the embedding network in orderto learn the variation between the intra-class. By increasing the network depth of the embedding module, we are able to achieve competitive performance due to the minimized intra-class variation.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.11 no.1 s.53
• /
• pp.59-65
• /
• 2007

Site classification of IBC and KBC is based on the ft-kips unit system and is not friendly for the linear interpolation of the site coefficients due to the implicit relationship between a site class and site coefficients, defining a site class by the range of the soil properties, not by a single soil property. Also, the site class definition of KBC has too wide range of soil properties for each soil class. making the structural engineers difficult to estimate the site coefficients for the diverse soil layers. In this study, a new site classification in SI unit system was proposed for the seismic design codes of KBC etc., and the comparison of the site coefficients of $F_{a}\;and\;F_{v}$ was also performed to investigate the possibility of the linear interpolation of the site coefficients with the proposed new site classification. According to the study results, it was more reasonable for the linear interpolation of the site coefficients to utilize the proposed new site classification considered the Sl unit system and the soil characteristics of the 30m soil layer beneath the shallow embedded foundation, and the linear interpolation of the acceleration coefficients for the design spectrum can be performed more reasonably defining the site coefficients for the representative shear wave velocities of each site class. With the study results, a new site classification, and the linear interpolation permitted acceleration coefficients fer the design spectrum were proposed for the modification of the seismic design regulations of KBC.

• Communications for Statistical Applications and Methods
• /
• v.26 no.6
• /
• pp.635-644
• /
• 2019

Class probability is a fundamental target in classification that contains complete classification information. In this article, we propose a class probability estimation method when the predictor is functional. Motivated by Wang et al. (Biometrika, 95, 149-167, 2007), our estimator is obtained by training a sequence of functional weighted support vector machines (FWSVM) with different weights, which can be justified by the Fisher consistency of the hinge loss. The proposed method can be extended to multiclass classification via pairwise coupling proposed by Wu et al. (Journal of Machine Learning Research, 5, 975-1005, 2004). The use of FWSVM makes our method model-free as well as computationally efficient due to the piecewise linearity of the FWSVM solutions as functions of the weight. Numerical investigation to both synthetic and real data show the advantageous performance of the proposed method.

• Journal of the Korean Data and Information Science Society
• /
• v.15 no.3
• /
• pp.565-574
• /
• 2004

Classification tree algorithms including as CART by Brieman et al.(1984) in some aspects, recursively partition the data space with the aim of making the distribution of the class variable as pure as within each partition and consist of several steps. SUPPORT(smoothed and unsmoothed piecewise-polynomial regression trees) method of Chaudhuri et al(1994), a weighted averaging technique is used to combine piecewise polynomial fits into a smooth one. We focus on applying SUPPORT to a binary class variable. Logistic model is considered in the caculation techniques and the results are shown good classification rates compared with other methods as CART, QUEST, and CHAID.

• Proceedings of the IEEK Conference
• /
• summer
• /
• pp.447-450
• /
• 2004

This paper proposes remotely sensed image classification method by fuzzy c-means clustering algorithm using average intra-cluster distance. The average intra-cluster distance acquires an average of the vector set belong to each cluster and proportionates to its size and density. We perform classification according to pixel's membership grade by cluster center of fuzzy c-means clustering using the mean-values of training data about each class. Fuzzy c-means algorithm considered membership degree for inter-cluster of each class. And then, we validate degree of overlap between clusters. A pixel which has a high degree of overlap applies to the maximum likelihood classification method. Finally, we decide category by comparing with fuzzy membership degree and likelihood rate. The proposed method is applied to IKONOS remote sensing satellite image for the verifying test.

• The Journal of the Korean dental association
• /
• v.15 no.2
• /
• pp.107-110
• /
• 1977

Class I malocclusion is essentially a dental dysplasia. Rotations, individual tooth malpositions, missing teeth, tooth size discrepancies, etc., fall under this classification. There are two types of class I malocclusions. One is identified by and insufficient denture base to accommodate the teeth; the other has more denture base than tooth material, creating spaces in the arch. The tooth material-to denture base discrepancies may be slight, calling for only a little increase in arch length for alignment and the correction of minor rotations. Discrepancies may also be great, in which case it becomes necessary to reduce tooth material by extraction, so as to make the tooth material more in proportion to the size of the denture base. The author had attempted orthodontic treatment of a class I malocclusion case of 13-year old boy in which high canines and impacted mandibular second premolars were involved. The author obtained good results.

• Proceedings of the IEEK Conference
• /
• 2004.06a
• /
• pp.179-182
• /
• 2004

The process of categorizing packets into 'class' in an Internet router is called packet classification. All packets with same class obey predefined rule specified in routing tables. Performing classification in real time on an arbitrary number of fields is a very challenge task. In this paper, we present a new algorithm named EnBiT-PC (EnBiT Packet Classification). and evaluate its performance against real classifiers in use today. We compare with previous algorithms, and found out that EnBiT-PC classify packets very efficiently and has relatively small storage requirements.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.619-622
• /
• 2001

In this paper, an efficient approach to classification of transient and harmonic disturbances in power systems is proposed. First, the Stop-and-Go CA CFAR Detector is utilized to detect a disturbance from the power signals which are mixed with other disturbances and noise. Then, (i) Wigner Distribution, SVD(Singular Value Decomposition) and Fisher´s Criterion (ii) DWT and Fisher´s Criterion, are applied to extract an efficient feature vector. For the classification procedure, a combined neural network classifier is proposed to classify each corresponding disturbance class. Finally, the 10 class data simulated by Matlab power system blockset are used to demonstrate the performance of the proposed classification system.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.33 no.4
• /
• pp.267-275
• /
• 2015

Those land cover maps have widely been used in various fields, such as environmental studies, military strategies as well as in decision-makings. This study proposes a method to extract training data, automatically and classify the cover using ingle satellite images and national land cover maps, provided by the Ministry of Environment. For this purpose, as the initial training data, those three were used; the unsupervised classification, the ISODATA, and the existing land cover maps. The class was classified and named automatically using the class information in the existing land cover maps to overcome the difficulty in selecting classification by each class and in naming class by the unsupervised classification; so as achieve difficulty in selecting the training data in supervised classification. The extracted initial training data were utilized as the training data of MLC for the land cover classification of target satellite images, which increase the accuracy of unsupervised classification. Finally, the land cover maps could be extracted from updated training data that has been applied by an iterative method. Also, in order to reduce salt and pepper occurring in the pixel classification method, the MRF was applied in each repeated phase to enhance the accuracy of classification. It was verified quantitatively and visually that the proposed method could effectively generate the land cover maps.