• Journal of Internet Computing and Services
• /
• v.15 no.3
• /
• pp.109-116
• /
• 2014

This study proposed new instance selection using neural network with weighted fuzzy membership functions(NEWFM) based on Takagi-Sugeno(T-S) fuzzy model to improve the classification performance. The proposed instance selection adopted weighted average defuzzification of the T-S fuzzy model and an interval selection, same as the confidence interval in a normal distribution used in statistics. In order to evaluate the classification performance of the proposed instance selection, the results were compared with depending on whether to use instance selection from the case study. The classification performances of depending on whether to use instance selection show 77.33% and 78.19%, respectively. Also, to show the difference between the classification performance of depending on whether to use instance selection, a statistics methodology, McNemar test, was used. The test results showed that the instance selection was superior to no instance selection as the significance level was lower than 0.05.

• Proceedings of the KAIS Fall Conference
• /
• 2002.11a
• /
• pp.205-208
• /
• 2002

본 연구에서는 영문 단어로부터 폰트를 분류하기 위해 연역적이고 국부적인 폰트 분류 방법을 제안한다. 이는 문자 인식 전에 한 단어에서 폰트를 분류하는 것을 말한다. 폰트 분류를 위해 활자 특성인 Ascender, Descender와 Serif가 사용된다. 입력 단어로부터 Ascender, Descender와 Serif가 추출되어 특징 벡터가 추출되고, 그 특징 벡터는 인공 신경망에 의해 입력 단어에 대한 폰트 그룹, 폰트 이름이 분류된다. 제안된 연역적이고 국부적인 폰트 분류 방법은 폰트 정보가 문자 분할기와 문자 인식기에 사용될 수 있게 한다 나아가, 특정 폰트에 따른 Mono-font 문자 분할기와 Mono-Font 문자 인식기로 구성되는 OCR 시스템을 구성할 수 있는 것을 가능하게 한다.

• Journal of the Society of Disaster Information
• /
• v.19 no.1
• /
• pp.44-59
• /
• 2023

Purpose: In the event of mass casualties, triage must be done promptly and accurately so that as many patients as possible can be recovered and returned to the battlefield. However, medical personnel have received many tasks with less manpower, and the battlefield for classifying patients is too complex and uncertain. Therefore, we studied an artificial intelligence model that can assist and replace medical personnel on the battlefield. Method: The triage model is presented using reinforcement learning, a field of artificial intelligence. The learning of the model is conducted to find a policy that allows as many patients as possible to be treated, taking into account the condition of randomly set patients and the medical capability of the military hospital. Result: Whether the reinforcement learning model progressed well was confirmed through statistical graphs such as cumulative reward values. In addition, it was confirmed through the number of survivors whether the triage of the learned model was accurate. As a result of comparing the performance with the rule-based model, the reinforcement learning model was able to rescue 10% more patients than the rule-based model. Conclusion: Through this study, it was found that the triage model using reinforcement learning can be used as an alternative to assisting and replacing triage decision-making of medical personnel in the case of mass casualties.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.3
• /
• pp.413-420
• /
• 2023

For a long time, researchers have presented excellent results in the field of image retrieval due to many studies on CBIR. However, there is still a semantic gap between these search results for images and human perception. It is still a difficult problem to classify images with a level of human perception using a small number of images. Therefore, this paper proposes an image classification model using deep learning-based transfer learning to minimize the semantic gap between images of people and search systems in image retrieval. As a result of the experiment, the loss rate of the learning model was 0.2451% and the accuracy was 0.8922%. The implementation of the proposed image classification method was able to achieve the desired goal. And in deep learning, it was confirmed that the CNN's transfer learning model method was effective in creating an image database by adding new data.

• Review of KIISC
• /
• v.12 no.6
• /
• pp.1-11
• /
• 2002

지금까지의 전쟁은 주로 국가 및 군사 분야에 한정하는 경향이 있어왔으나 정보전은 기업은 물론이고 일반 개인에게까지 직접 관련이 되고 있다. 그러나 아직 까지 정보전에 대한 명확한 개념이 정립되지 않은 상태에서 정보전과 관련된 용어들이 난무하고 있어 일반인은 말할 것도 없고 정책 입안자나 군사 기획가마저도 많은 혼란을 겪고 있다. 따라서 본 고에서는 이러한 정보전에 대한 명확한 이해를 도모할 목적으로 정보전이 출현된 배경, 정보전의 정의 및 분류, 정보전의 특징 등을 살펴보고 몇 가지 쟁점들에 대해 논한다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.655-657
• /
• 2021

It is challenging to find breast ultrasound image training dataset to develop an accurate machine learning model due to various regulations, personal information issues, and expensiveness of acquiring the images. However, studies targeting transfer learning for ultrasound breast cancer images classification have not been able to achieve high performance compared to radiologists. Here, we propose an improved transfer learning model for ultrasound breast cancer classification using publicly available dataset. We argue that with a proper combination of ImageNet pre-trained model and optimizer, a better performing model for ultrasound breast cancer image classification can be achieved. The proposed model provided a preliminary test accuracy of 99.5%. With more experiments involving various hyperparameters, the model is expected to achieve higher performance when subjected to new instances.

• 방재와보험
• /
• s.116
• /
• pp.70-71
• /
• 2006

사과에도 중심(core)이 있고 지구도 지표면과 안으로 중심이 있듯이 인체에도 몸의 중심 부위가 있다. 보통 단전 부위에 무게중심이 있다고 설명을 한다. 단전의 위치에 대해서는 여러 가지 설이 있어 한마디로 정의하기는 어려우나 일반적으로 상단전, 중단전, 하단전으로 분류학 상단전은 뇌 부분, 중단전은 심장에서 명치 부분, 하단전은 배꼽 아래 부분에 있다고 한다. 단전이라고 할 때 약간의 차이는 있지만 아랫배 부근을 가르키며, 기해라고도 한다. 특히 하단전은 모든 경락이 모이는 곳으로서 원기를 저장하는 곳이며 기 흐름의 요체이다. 또한 생명력을 배양하는 곳이자 복식호흡의 기본력이라고 알려져 있다. 그러면 의학적으로는 어떤 것을 신체 중심(core)이라고 이야기 할까?

• The Journal of the Korea Contents Association
• /
• v.9 no.3
• /
• pp.305-311
• /
• 2009

The bankruptcy in Korea affects to all stakeholder of firms. Companies listed in KOSDAQ have high technology but the possibilities for success of business are low. The purpose of this study is to develop and to applicate falling prediction model of KOSDAQ firms using logistic regression analysis. The results of this study are as follows. First, the accuracy of classification of the models by years was between 76.5% and 77.5%, and that of the mean model was between 70.6% and 83.4%. Among the models, the mean model of -three years, -two years, and -one year was highest in the accuracy of classification (83.4%). Second, when the mean model of -three year, -two years, and -one years, the highest model in accuracy of classification, was selected to be verified on validation samples, the accuracy of prediction increased from -three years to -one year (71.7% for -three years, 75.0% for -two years, 90.0% for -one year). In indicating the superiority of developed model.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.27 no.6
• /
• pp.814-824
• /
• 1994

The taxonomic revision of the family Mugilidae from Korea was made based on the fish specimens collected from the coasts of the Korea from July 1990 to July 1994. The family Mugilidae was classified into three species belonging to two genera: Mugil cephalus, Liza haematocheilus and Liza carinatus. Previousely Mugil japonicus reported as a species from Korea was confirmed into junior synonym of Mugil cephalus. based on the external and internal morphological characters. A new key to the genera and species of family Mugilidae was proposed and described their distribution in Korea.

• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10b
• /
• pp.154-156
• /
• 1999

문서 분류는 미리 정의된 두 개 또는 그 이상의 클래스에 새로 생성되는 객체들을 할당하는 방법이다. 문서의 자동 분류에 대한 연구는 오래 전부터 연구되어 왔지만 한국어에 대한 적용 및 연구는 다른 분야에 비해 아직까지 활발히 이루어지지 않고 있다. 본 논문에서는 문서를 자동으로 분류하기 위해 문서의 주제어에 가중치를 부여하고, 부족한 문서의 특징을 보충하기 위하여 말뭉치로부터 주제어들과의 상호정보에 의해 추출된 단어를 사용하여 문서를 표현한 후, 가중치를 부여한 문서의 주제어에 베이지안 분류자를 사용하여 문서분류를 수행한다. 실험은 한국어 정보검색 실험용 데이터 집합인 KTset95 문서 4,414개 중 1,300개의 문서를 학습 집합으로, 1,000개의 문서를 분류에 대한 검증 집합으로 사용하였다. 실험 결과, 순수 베이지안 확률을 사용한 기존의 방법보다 실험 집합과 검증 집합에서 각각 1.92%, 4.3% 향상된 분류 정확도를 얻었다.