• 전자공학회논문지B
• /
• 제30B권2호
• /
• pp.8-18
• /
• 1993

A multi-font, multi-size and high speed character recognition system is designed. The design principles are simpilcity of algorithm, adaptibility, learnability, hierachical data processing and attention by feed back. For the multi-size character recognition, the extracted character images are normalized. A hierachical classifier classifies the feature vectors. Feature is extracted by applying the directional receptive field after the directional dege filter processing. The hierachical classifier is consist of two pre-classifiers and one decision making classifier. The effect of two pre-classifiers is prediction to the final decision making classifier. With the pre-classifiers, the time to compute the distance of the final classifier is reduced. Recognition rate is 95% for the three documents printed in three kinds of fonts, total 1,700 characters. For high speed implemention, a multiprocessor system with the ring structure of four transputers is implemented, and the recognition speed of 30 characters per second is aquired.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 추계학술대회
• /
• pp.385-389
• /
• 2003

Three-dimensional experimental analysis was conducted in the pulverizer simplified isothermal model. The experiment model was constructed on a 1/3.5 scale of 500MW pulverizer. The purpose of this study is to investigate the effect of design parameters on the pulverized coal separator efficiency. Where used pulverized coal separator design parameters are guide vane angle, static classifier angle, dynamic classifier rpm. Taguchi method was used to find the effective design parameters related to pulverized coal separator efficiency. The results of the experiment showed that guide vane angle and dynamic classifier rpm were the design key parameters. In addition to the total number of experiment cases were reduced by Taguchi method.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 1998년도 Proceedings of International Symposium on Remote Sensing
• /
• pp.122-127
• /
• 1998

This paper presents the neuro-fuzzy classifier derived from the generic model of a 3-layer fuzzy perceptron and developed the classification software based on the neuro-fuzzl model. Also, a comparison of the neuro-fuzzy and maximum-likelihood classifiers is presented in this paper. The Airborne Multispectral Scanner(AMS) imagery of Tae-Duk Science Complex Town were used for this comparison. The neuro-fuzzy classifier was more considerably accurate in the mixed composition area like "bare soil" , "dried grass" and "coniferous tree", however, the "cement road" and "asphalt road" classified more correctly with the maximum-likelihood classifier than the neuro-fuzzy classifier. Thus, the neuro-fuzzy model can be used to classify the mixed composition area like the natural environment of korea peninsula. From this research we conclude that the neuro-fuzzy classifier was superior in suppression of mixed pixel classification errors, and more robust to training site heterogeneity and the use of class labels for land use that are mixtures of land cover signatures.

• 한국컴퓨터정보학회논문지
• /
• 제23권6호
• /
• pp.1-7
• /
• 2018

A multi-label classification is to find multiple labels associated with the input pattern. Multi-label classification can be achieved by extending conventional single-label classification. Common extension techniques are known as Binary relevance, Label powerset, and Classifier chains. However, most of the extended multi-label naive bayes classifier has not been able to accurately estimate posterior probabilities because it does not reflect the label dependency. And the remaining extended multi-label naive bayes classifier has a problem that it is unstable to estimate posterior probability according to the label selection order. To estimate posterior probability well, we propose a new posterior probability estimation method that reflects the probability between all labels and labels efficiently. The proposed method reflects the correlation between labels. And we have confirmed through experiments that the extended multi-label naive bayes classifier using the proposed method has higher accuracy then the existing multi-label naive bayes classifiers.

• 제어로봇시스템학회논문지
• /
• 제16권1호
• /
• pp.96-100
• /
• 2010

This paper proposes novel genetic design of optimal classifier ensemble for vehicle detection using Genetic Algorithm with Varying Population Size (GAVaPS). Recently, many classifiers are used in classifier ensemble to deal with tremendous amounts of data. However the problem has a exponential large search space due to the increasing the number of classifier pool. To solve this problem, we employ the GAVaPS which outperforms comparison with simple genetic algorithm (SGA). Experiments are performed to demonstrate the efficiency of the proposed method.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2000년도 춘계학술대회 학술발표 논문집
• /
• pp.179-182
• /
• 2000

A Classifier System processes a discrete coded information from the environment. When the system codes the information to discontinuous data, it loses excessively the information of the environment. The Fuzzy Classifier System(FCS) makes the classifier system be able to carry out the mapping from continuous inputs to outputs. It is the FCS that applies this ability of the machine learning to the concept of fuzzy controller. It is that the antecedent and consequent of classifier is same as a fuzzy rule of the rule base. In this paper, the FCS is the Michigan style and fuzzifies the input values to create the messages. The system stores those messages in the message list and uses the implicit Bucket Brigade Algorithms. Also the FCS employs the Genetic Algorithms(GAs) to make new rules and modify rules when performance of the system needs to be improved. We will verify the effectiveness of the proposed FCS by applying it to AMR avoiding the obstacle.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제5권3호
• /
• pp.263-267
• /
• 2005

Building classifiers for financial real-world classification problems is often plagued by severely overlapping and highly skewed class distribution. New performance measures such as receiver operating characteristic (ROC) curve and area under ROC curve (AUC) have been recently introduced in evaluating and building classifiers for those kind of problems. They are, however, in-effective to evaluation of classifier's discrimination performance in a particular class of the classification problems that interests lie in only a local operating range of the classifier, In this paper, a new method is proposed that enables us to directly improve classifier's discrimination performance at a desired local operating range by defining and optimizing a partial area under ROC curve or domain-specific curve, which is difficult to achieve with conventional classification accuracy based learning methods. The effectiveness of the proposed approach is demonstrated in terms of fraud detection capability in a real-world fraud detection problem compared with the MSE-based approach.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2008년도 추계학술대회B
• /
• pp.2423-2428
• /
• 2008

A turbo classifier having a rotating rotor of two stage classification region has been developed to have a cut size of 1 micro meter. Particle number concentrations were counted using Aerosol Particle Sizer (APS, TSI co., USA) at inlet and outlet of the classifier. Partial classification efficiency was obtained at various rotation speeds, total flow rates, and feed rates of powders, and classification characteristic depending on design parameters was discussed. Classification performance was enhanced as rotation speed of rotor increased and total flow rate decreased.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1995년도 추계학술대회 학술발표 논문집
• /
• pp.11-14
• /
• 1995

A Fuzzy Classifier which deals with very confusing objects is proposed. Naturally this classifier heavily relies on the nulti-feature decision-making procedure. For a simple example, this classifier is applied to the recognition of confusing handwritten numerals 4,6 and 9 The characteristic variables used in this paper are the existence of a loop and the relative location of the starting or ending points(SEP). Thus each sample of handwritten numerals 4, 6 and 9 is classified in one of the 6 groups which are divided according to the sample structure. Each group has its own classifying rules. Also the method of rule-generation using genetic algorithms in each group is proposed.

• 지능정보연구
• /
• 제3권2호
• /
• pp.83-91
• /
• 1997

DELVAUX is an inductive learning environment that learns Bayesian classification rules from a set o examples. In DELVAUX, a genetic a, pp.oach is employed to learn the best rule-set, in which a population consists of rule-sets and rule-sets generate offspring by exchanging some of their rules. We have explored a meta-learning a, pp.oach in the DELVAUX learning environment to improve the classification performance of the DELVAUX system. The meta-learning a, pp.oach learns the bias of a classifier so that it can evaluate the prediction made by the classifier for a given example and thereby improve the overall performance of a classifier system. The paper discusses the meta-learning a, pp.oach in details and presents some empirical results that show the improvement we can achieve with the meta-learning a, pp.oach.