• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.451-452
• /
• 2007

본 논문에서는 다항식 기반 Radial Basis Function(RBF)신경회로망(Polynomial based Radial Basis Function Neural Networks)을 설계하고 이를 2-클래스 패턴 분류 문제에 응용하여 그 성능을 분석한다. 제안된 다항식기반 RBF 신경회로망은 입력층, 은닉층, 출력 층으로 이루어진다. 입력층은 입력 벡터의 값들을 은닉 층으로 전달하는 기능을 수행하고 은닉층은 Fuzzy c-means 클러스터링을 통하여 뉴런의 출력 값으로 내보낸다. 은닉층과 출력층사이의 연결가중치는 상수, 선형식 또는 이차식으로 이루어지며 경사 하강법에 의해 학습된다. Networks의 최종 출력은 연결가중치와 은닉층 출력의 곱에 의해 퍼지추론의 결과로서 얻어진다. 제안된 다항식기반 RBF 신경회로망은 각기 다른 4종류의 2-클래스 분류 문제에 적용 및 평가되어 분류기로써의 성능을 분석한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.4 no.3
• /
• pp.611-618
• /
• 2000

This paper presents a shape classification method of dynamic image based on adaptive fuzzy inference. It describes the design scheme of fuzzy inference algorithm which makes it suitable for low speed systems such as conveyor, uninhabited transportation. In the first Discrete Wavelet Transform(DWT) is utilized to extract the motion vector in a sequential images. This approach provides a mechanism to simple but robust information which is desirable when dealing with an unknown environment. By using feature parameters of moving object, fuzzy if - then rule which can be able to adapt the variation of circumstances is devised. Then applying the implication function, shape classification processes are performed. Experimental results are presented to testify the performance and applicability of the proposed algorithm.

• Proceedings of the KIEE Conference
• /
• 1987.07b
• /
• pp.1157-1159
• /
• 1987

A syntactic and semantic approach is used to make type classification based on feature points(whorl, delta, core) and the shape of flow line around feature points. The image is divided into 30 by 30 subregions which are represented in the average direction and 4-tuple direction component. Next the relaxation process with singularity detection and convergency checking is performed. A set of semantic languages is used to describe the major flow line around the extracted feature points. LR(1) parser and feature transfer function are used to recognize the coded flow patterns. The 72 fingerprint impressions is used to test the proposed approach and the rate of the classification is about 93 percentages.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2000.04a
• /
• pp.601-604
• /
• 2000

In this paper, we compare the classification performance of both data fusion and clustering algorithms (Data Bagging, Variable Selection Bagging, Parameter Combining, Clustering) to logistic regression in consideration of various characteristics of input data. Four factors used to simulate the logistic model are (1) correlation among input variables (2) variance of observation (3) training data size and (4) input-output function. Since the relationship between input & output is not typically known, we use Taguchi design to improve the practicality of our study results by letting it as a noise factor. Experimental study results indicate the following: Clustering based logistic regression turns out to provide the highest classification accuracy when input variables are weakly correlated and the variance of data is high. When there is high correlation among input variables, variable bagging performs better than logistic regression. When there is strong correlation among input variables and high variance between observations, bagging appears to be marginally better than logistic regression but was not significant.

• Proceedings of the IEEK Conference
• /
• 2001.06b
• /
• pp.309-312
• /
• 2001

A primitive gas classification system which can classify limited species of gas was designed and simulated. The 'electronic nose' consists of an array of 4 metal oxide gas sensors with different selectivity patterns, signal collecting unit and a signal pattern recognition and decision Part in PLD(programmable logic device) chip. Sensor array consists of four commercial, tin oxide based, semiconductor type gas sensors. BP(back propagation) neutral networks with MLP(Multilayer Perceptron) structure was designed and implemented on CPLD of fifty thousand gate level chip by VHDL language for processing the input signals from 4 gas sensors and qualification of gases in air. The network contained four input units, one hidden layer with 4 neurons and output with 4 regular neurons. The 'electronic nose' system was successfully classified 4 kinds of industrial gases in computer simulation.

• International Journal of CAD/CAM
• /
• v.8 no.1
• /
• pp.69-72
• /
• 2009

In this paper, we present a new shape descriptor, which is named Poisson equation-Fourier-Mellin moment Descriptor. We solve the Poisson equation in the shape area, and use the solution to get feature function, which are then integrated using Fourier-Mellin moment to represent the shape. This method develops the Poisson equation-geometric moment Descriptor proposed by Lena Gorelick, and keeps both advantages of Poisson equation-geometric moment and Fourier-Mellin moment. It is proved better than Poisson equation-geometric moment Descriptor in shape recognition and classification experiments.

• Journal of the Korean Mathematical Society
• /
• v.59 no.2
• /
• pp.217-233
• /
• 2022

Although machine learning shows state-of-the-art performance in a variety of fields, it is short a theoretical understanding of how machine learning works. Recently, theoretical approaches are actively being studied, and there are results for one of them, margin and its distribution. In this paper, especially we focused on the role of margin in the perturbations of inputs and parameters. We show a generalization bound for two cases, a linear model for binary classification and neural networks for multi-classification, when the inputs have normal distributed random noises. The additional generalization term caused by random noises is related to margin and exponentially inversely proportional to the noise level for binary classification. And in neural networks, the additional generalization term depends on (input dimension) × (norms of input and weights). For these results, we used the PAC-Bayesian framework. This paper is considering random noises and margin together, and it will be helpful to a better understanding of model sensitivity and the construction of robust generalization.

• Journal of Korean Academy of Nursing Administration
• /
• v.7 no.1
• /
• pp.15-23
• /
• 2001

This study was a methodological research to develop computerized patient classification system. The subjects of this investigation were 435 inpatients except redundant data and outliers in P University Hospital from January 18, 2000 to January 24, 2000. The data was analyzed by discrimination analysis and adopted discriminant variables were 1) sum of frequency for the nursing activities, 2) the number of nursing activities that do not need to consider intensity of the activities, and 3) total hours of nursing activities that need to consider their intensities. Discriminant function developed by this study classified the patients into 4 groups; class I, 251 ; class II, 125 ; class III, 39 ; class IV, 20. The Hit ratio was 89.23. Based on this study, following suggestions can be made for the future research 1. Inclusive patient classification system, which includes more expanded direct nursing care factors, need to be developed and examined. 2. This developed classification system can be utilized to evaluate patient distribution and to estimate adequate numbers of nursing staffs in each nursing unit.

• The Korean Journal of Applied Statistics
• /
• v.27 no.5
• /
• pp.773-786
• /
• 2014

We propose the multiplication of false rates (MFR) which is a classification accuracy criteria and an area type of rectangle from ROC curve. Optimal threshold obtained using MFR is compared with other criteria in terms of classification performance. Their optimal thresholds for various distribution functions are also found; consequently, some properties and advantages of MFR are discussed by comparing FNR and FPR corresponding to optimal thresholds. Based on general cost function, cost ratios of optimal thresholds are computed using various classification criteria. The cost ratios for cost curves are observed so that the advantages of MFR are explored. Furthermore, the de nition of MFR is extended to multi-dimensional ROC analysis and the relations of classification criteria are also discussed.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.842-857
• /
• 2023

Currently, most sentiment classification models on microblogging platforms analyze sentence parts of speech and emoticons without comprehending users' emotional inclinations and grasping moral nuances. This study proposes a hybrid sentiment analysis model. Given the distinct nature of microblog comments, the model employs a combined stop-word list and word2vec for word vectorization. To mitigate local information loss, the TextCNN model, devoid of pooling layers, is employed for local feature extraction, while BiLSTM is utilized for contextual feature extraction in deep learning. Subsequently, microblog comment sentiments are categorized using a classification layer. Given the binary classification task at the output layer and the numerous hidden layers within BiLSTM, the Tanh activation function is adopted in this model. Experimental findings demonstrate that the enhanced TextCNN-BiLSTM model attains a precision of 94.75%. This represents a 1.21%, 1.25%, and 1.25% enhancement in precision, recall, and F1 values, respectively, in comparison to the individual deep learning models TextCNN. Furthermore, it outperforms BiLSTM by 0.78%, 0.9%, and 0.9% in precision, recall, and F1 values.