• Journal of Korea Society of Industrial Information Systems
• /
• v.27 no.1
• /
• pp.49-62
• /
• 2022

Online customer review data can be easily collected on the Internet and also they describe sentimental evaluation of a product in different aspects. Previous sentiment analysis studies evaluate the degree of sentiment with review data, which may have multiple sentences describing different product aspects. Since different aspects of a product can be described in a sentence, the proposed method suggested analyzing a sentence to build a pair of a product aspect terms and sentimental terms. Bidirectional LSTM and CRF algorithms were used in this paper. A pair of aspect terms and sentimental terms are evaluated by pre-defined evaluation rules. The paper suggested using the result of evaulation as inputs of QFD, so that the quantified customer voices effect on the requirements of a new product. Online reviews for a hair dryer were used as an example showing that the proposed approach can derive reasonable sentiment analysis results.

• Journal of Digital Convergence
• /
• v.12 no.6
• /
• pp.223-229
• /
• 2014

In an attempt to retrieve useful information for an efficient decision in the large knowledge system, it is generally necessary and important for a refined feature selection. Rough set has difficulty in generating optimal reducts and classifying boundary objects. In this paper, we propose quick reduction algorithm generating optimal features by rough entropy analysis for condition and decision attributes to improve these restrictions. We define a new conditional information entropy for efficient feature extraction and describe procedure of feature selection to classify the significance of features. Through the simulation of 5 datasets from UCI storage, we compare our feature selection approach based on rough set theory with the other selection theories. As the result, our modeling method is more efficient than the previous theories in classification accuracy for feature selection.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1995.05a
• /
• pp.841-846
• /
• 1995

대전지역 토양에서 추출한 흄산(TJHA)과 Am(III), Eu(III)의 착물형성에 대한 안정도상수를 추출용매 di-2-ethylhexyl phosphoric acid와 희석제 toluene을 사용하여 용매추출법으로 구하였다. 이온강도가 0.1M NaCIO$_4$에서 TJHA의 총 carboxylate capacity를 직접 전위차적정법으로 분석한 결과 3.757 meq/g이고, apparent pKa는 5.15 이었다. TJHA와 Am(III) 및 Eu(III)의 조건부 안정도상수의 log$\beta$1 값과 log$\beta$2 값을 흄산의 이온화도 함수로 구한 결과, Eu-TJHA은 0.1M NaClO$_4$일때 log $\beta$1=5.948$\alpha$ + (6.83 $\pm$ 0.3) 및 log $\beta$2 = 5.687$\alpha$ + (10.44 $\pm$ 0.4)이며, Am-TJHA은 log$\beta$$_1$= 4.004 $\alpha$ + (6.96 $\pm$ 0.2) 및 log$\beta$$_2$= 3.719 $\alpha$ + (11.71 $\pm$ 0.2)이었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.259-261
• /
• 2004

생물학 분야의 문헌으로부터 유전자 및 단백질 이름을 추출하는 기술은 바이오 텍스트 마이닝 분야의 기반 기술로 그 중요성이 점차 증대되고 있다. 이 연구에서는 생물학 분야 문헌의 초록으로부터 하나의 토큰으로 구성된 single gene name은 물론 여러 개의 토큰으로 이루어진 multi gene name까지 유전자나 단백질의 이름을 자동으로 추출하는 시스템 TagGeN(Tagger for Gene Name)을 제안한다. TagGeN은 기존의 태거와 달리, 문자나 숫자 이외의 기호를 포함한 유전자나 단백질 이름의 품사 처리에 있어 개선 방안을 제시하고, 여러 토큰으로 이루어진 이름의 인식에 있어 나란한 두 토큰이 갖는 태그 정보를 이용한 조건부 확률에 근거하여 Markov 모델을 도입한다. 위와 같은 개선방안을 구현한 TagGeN은 성능면에서 기존의 유사시스템에 비해 recall 20.8%, precision 4.7%의 성능향상을 보임으로써 본 연구에서 제안한 방법론의 효과를 입증한다.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.172-175
• /
• 2018

환자의 병력을 서술하는 임상문서에서 임상 개체명들은 그들 사이에 개체명이 아닌 단어들이 위치하기 때문에 거리상으로 서로 떨어져 있고, 임상 개체명인식에 많이 사용되는 조건부무작위장(conditional random fields; CRF) 모델은 Markov 속성을 따르기 때문에 서로 떨어져 있는 개체명 라벨 간의 전이 정보는 모델의 계산에서 무시된다. 본 논문에서는 라벨링 모델에 서로 떨어진 개체명 간 전이 관계를 표현하기 위하여 CRF 모델의 구조를 변경하는 방법론을 소개한다. 제안된 CRF 모델 디자인에서는 모델의 계산효율성을 빠르게 유지하기 위하여 Markov 속성을 유지하는 1차 모델 구조를 유지한다. 모델은 선행하는 개체명의 라벨 정보를 후행하는 개체명 엔터티에게 전달하기 위하여 선행 개체명의 라벨을 뒤 따르는 비개체명 라벨에 전이시키고 이를 통해 후행하는 개체명은 선행하는 개체명의 라벨 정보를 알 수 있게 된다. 라벨의 고차 전이 정보를 전달함에도 모델의 구조는 1차 전이 구조를 유지함으로 n차 구조의 모델보다 빠른 계산 속도를 유지할 수 있게 된다. 모델의 성능 평가를 위하여 서울대학교병원 류머티즘내과에서 퇴원한 환자들의 퇴원요약지에 병력과 관련된 엔터티가 태깅된 평가 데이터와 i2b2 2012/VA 임상자연어처리 shared task의 임상 개체명 추출 데이터를 사용하였고 기본 CRF 모델들(1차, 2차)과 비교하였다. 피처 조합에 따라 모델들을 평가한 결과 제안한 모델이 거의 모든 경우에서 기본 모델들에 비하여 F1-score의 성능을 향상시킴을 관찰할 수 있었다.

• Journal of Korea Water Resources Association
• /
• v.43 no.9
• /
• pp.835-842
• /
• 2010

This paper attempts to examine households' willingness to pay (WTP) for ground water pollution improvement which can be used in cost-benefit analysis on the project for developing the soil pollution control technique. We applied a contingent valuation (CV) method to obtain at least a preliminary evaluation of the WTP. The CV survey was rigorously designed to comply with the guidelines for best-practiced CV studies. We surveyed a randomly selected sample of 500 households in Seoul metropolitan area and asked respondents questions in person-to-person interviews about how they would be willing to pay for the program. Respondents overall accepted the contingent market and were willing to contribute a significant amount (1,195 to 1,552 won), on average, per household per year. The aggregate value of the project for developing the soil pollution control technique amounts to approximately 20.3 billion won per year. The household values can be the benefits that ensue from the project and compared with the costs of the program to determine whether the project is economically desirable.

• Environmental and Resource Economics Review
• /
• v.16 no.3
• /
• pp.451-483
• /
• 2007

Recently, the problem of noise has received much attention in the urban environment. This paper attempts to measure the economic benefits of urban noise reduction in the metropolitan area. To this end, the dichotomous choice contingent valuation method is applied. In particular, recently proposed one and one-half bound model that reduces the potential for response bias in the double bound model while maintaining much of its efficiency. We surveyed a randomly selected sample of 800 households in the metropolitan area and asked respondents questions in person-to-person interviews about how they would willing to pay for the noise reduction. Respondents overall accepted the contingent market and were willing to contribute a significant amount (997 to 1,778 won), on average, per household per month. This willingness varies according to individual characteristics such as concerns about noise, dwelling area, and income. The aggregate value of the noise reduction in the sampled metropolitan area amounts to approximately 79.26 to 141.35 billion won per year.

• Journal of Environmental Policy
• /
• v.5 no.3
• /
• pp.79-103
• /
• 2006

This paper attempts to examine household's willingness to pay (WTP) for piped water quality improvement in Wonju, where the local government are planning to implement a piped water quality improvement program. We apply a contingent valuation (CV) method to obtain at least a preliminary evaluation of the WTP. The CV survey was rigorously designed to comply with the guidelines for best-practiced CV studies. We surveyed a randomly selected sample of 250 households in Wonju and asked respondents questions in person-to-person interviews about how much they would be willing to pay for the program. Respondents overall accepted the contingent market and were willing to contribute a significant amount (1,583 to 2,776 won), on average, per household per month. This willingness varies according to individual characteristics such as sex, education level, and income. The aggregate value of the program in Wonju amounts to approximately 1.99 billion won to 3.49 billion won per year. The household values can be the benefits that ensue from the program and compared with the costs of the program to determine whether the program is economically desirable.

• Journal of Korea Society of Industrial Information Systems
• /
• v.4 no.2
• /
• pp.1-6
• /
• 1999

We propose a new algorithm for pattern classification by extracting feature vectors based on Gibbs distributions which are well suited for representing the characteristic of an images. The extracted feature vectors are comprised of 2-D moments which are invariant under translation rotation, and scale of the image less sensitive to noise. This implementation contains two puts: feature extraction and pattern classification First of all, we extract feature vector which consists of an improved 2-D moments on the basis of estimated Gibbs distribution Next, in the classification phase the minimization of the discrimination cost function for a specific pattern determines the corresponding template pattern. In order to evaluate the performance of the proposed scheme, classification experiments with training document sets of characters have been carried out on SUN ULTRA 10 Workstation Experiment results reveal that the proposed scheme had high classification rate over 98%.

• Proceedings of the KIEE Conference
• /
• 2015.07a
• /
• pp.1355-1356
• /
• 2015

본 논문에서는 Convolution기법 기반 RBFNNs 패턴 분류기를 사용한 2차원 얼굴인식 시스템을 설계한다. 제안된 방법은 특징 추출과 차원축소를 하는 컨볼루션 계층과 부분추출 계층을 교대로 연결하여 2차원 이미지를 1차원의 특징 배열로 만든다. 그 후, 만들어진 1차원의 특징 배열을 RBFNNs 패턴 분류기의 입력으로 사용하여 인식을 수행한다. RBFNNs의 조건부에는 FCM 클러스터링 알고리즘을 사용하며 연결가중치는 1차 선형식을 사용하였다. 또한 최소 자승법(LSE : Least Square Estimation)을 사용하여 다항식의 계수를 추정하였다. 제안된 모델의 성능을 평가하기 위해 CMU PIE Database를 사용한다.