• Communications for Statistical Applications and Methods
• /
• 제6권3호
• /
• pp.831-842
• /
• 1999

We propose a wavelet decomposition feature extraction method for the hand-written character recognition. Comparing the recognition rates of which methods with original image features and with selected features by the wavelet decomposition we study the characteristics of the proposed method. LDA(Linear Discriminant Analysis) QDA(Quadratic Discriminant Analysis) RDA(Regularized Discriminant Analysis) and NN(Neural network) are used for the calculation of recognition rates. 6000 hand-written numerals from CENPARMI at Concordia University are used for the experiment. We found that the set of significantly selected wavelet decomposed features generates higher recognition rate than the original image features.

• 한국축산식품학회지
• /
• 제35권4호
• /
• pp.494-501
• /
• 2015

The aim of this study was to examine the effectiveness of pre-cooking conditions on the quality characteristics of ready-to-eat (RTE) Samgyetang. Raw chickens were steamed under the different conditions of 50℃/30 min (T1), 65℃/30 min (T2), 85℃/30 min (T3), and 90℃/10 min (T4) prior to retorting at 120℃ for 65 min. The results showed that pre-cooking conditions in all treated samples could reduce fat contents in breast and leg meats by 8.5-11.7％ and 10.0-11.0％ compared to the control, even though there were no significant differences among treatments (p>0.05). The L* and b* values of breast and leg meats treated with the higher temperature and longer time conditions were significantly higher than the control (p<0.05), while a* values tended to decrease despite of not to a significant extent (p>0.05). Moreover, apparent viscosity and water soluble protein showed insignificant differences (p>0.05) among the samples as a result of the retorting process, which might have more negative influences on the quality. T2 samples obtained significantly the highest average Quantitative Descriptive Analysis (QDA) score and transmittance value, representing the most clear broth among the samples, compared to the control. On the other hand, T3 showed the highest cooking loss among the treatments and the lowest QDA scores among the samples. In conclusion, pre-cooking treatment prior to retorting in manufacturing Samgyetang is a plausible way to reduce its fat content. A pre-cooking condition at either 65℃ for 30 min, or 90℃ for 10 min are recommended for producing Samgyetang with optimum quality.

• 한국수산과학회지
• /
• 제39권6호
• /
• pp.431-436
• /
• 2006

This study was red snow crab Chionoecetes japonicus cooker effluent (RSCCE) for making a natural crab-like flavorant. The RSCCE ($1\;^{\circ}Brix$ in the initial state) was concentrated up to $40^{\circ}Brix$ to determine the optimal conditions for making a natural flavorant. During concentration, the amino-N content and total acidity increased with the concentration time, while the pH was maintained in range 7.94-8.78. In the acceptance test and quantitative description analysis (QDA), $20^{\circ}Brix$ RSCCE had the best quality in terms of taste (5.87), odor (6.00), and overall acceptance (5.80). Of the taste compounds analyzed in $20^{\circ}Brix$ RSCCE, lactic acid was an abundant non-volatile organic acid, and the nucleotide 5'-inosine monophosphate (IMP) was present, as were four free amino acids: tyrosine, glutamic acid, alanine and glycine. The taste and odor of boiled crabmeat were retained in $20^{\circ}Brix$ RSCCE based on the QDA.

• 한국식생활문화학회지
• /
• 제33권2호
• /
• pp.142-148
• /
• 2018

In this study, Seolgi Tteok was made in order to increase consumption of Geuk, which possesses numerous nutritional advantages, and its optimum content as a new food coloring for rice cake was determined. Quality characteristics and quantitative description analysis (QDA) of Seolgi Tteok, in which Geuk was added at 0, 4, 8, and 12%, was conducted. Consumer acceptability test was also conducted. As the amount of added Geuk increased, moisture content of Seolgi Tteok increased as well. Brightness (L-value) was the highest in the control group, and more Geuk resulted in higher values of red index (a-value) and yellow index (b-value), which corresponded to the results of the sensory evaluation. As result of the mechanical texture measurement, only adhesiveness and resilience show a significant difference. As a result of the QDA, 17 sensory characteristic terms were assessed. Among them, only 13 showed a significant difference. Among the different sensory characteristics, almost all of them except for taste characteristics were significantly influenced by the amount of Geuk. The aroma and taste of Geuk were not largely influenced. It can be suggested that Geuk is not a factor that strongly influences flavor. In conclusion, Geuk does not have a strong influence on the taste or aroma of Seolgi Tteok but does on color characteristics. It can be suggested that Geuk is qualified as a coloring material for food, and the reasonable addition amount is 8%. As a result of this research, Geuk can be considered as a coloring material for other types of rice cake, traditional Korean sweets, and even confectioneries as well as for Seolgi Tteok. This implies that Geuk can be utilized to develop various new products as a coloring material with abundant nutritional content, which will contribute to the promotion of Geuk consumption.

• 한국식생활문화학회지
• /
• 제2권1호
• /
• pp.71-79
• /
• 1987

한과류의 관능적 품질특성을 조사하기 위하여 약과, 세반강정, 산자, 송화다식, 들깨엿강정에 대한 품질표현용어를 77명의 설문자를 대상으로 조사하였으며 이 결과 90여종의 표현용어를 수집하였다. 이들 용어중 58종의 표현용어는 국어대사전에 수록되어 있었으며 한과의 종류별 주요품질요소를 표현빈도수에 근거하여 결정하였다. 한과류를 실온에서 10일간 상대습도 $0{\sim}68%$ 범위에서 저장하였을 때 주요품질요소의 변화를 관능검사로 평가하였다. 저장상대습도의 변화에 따른 한과류 종류별 주요 관능적 품질요소의 변화 양상을 정량적 묘사 분석법으로 도해함으로서 일목요연하게 표시할 수 있었다.

• Journal of the Korean Data Analysis Society
• /
• 제20권6호
• /
• pp.2853-2864
• /
• 2018

판별분석(discriminant analysis)은 새로운 개체가 입력되었을 때, 그 개체가 어느 그룹에 속하는지 예측하는데 사용되는 분석방법이다. 판별분석에서는 레이블(label)을 통해 새로운 개체를 예측하기 때문에 판별분석에서 레이블은 중요하다. 레이블 노이즈(label noise)는 관측된 레이블에 오류가 포함된 것을 의미하며, 실데이터에 발생하기 쉽고 판별성능에 영향을 미칠 수 있는 중요한 요인이다. 이를 개선하기 위해 레이블 노이즈와 레이블 노이즈에 강건한 모형들이 연구되고 있지만, 레이블 노이즈가 존재할 때 판별성능에 영향을 줄 수 있는 요인을 고려하고 이 요인들이 판별성능에 미치는 영향을 비교한 연구는 찾기 힘들다. 따라서 이 논문에서는 분류문제에서 많이 사용되는 LDA, QDA, KNN, SVM 방법을 이용하여 레이블 노이즈가 판별성능에 미치는 영향을 알아보고자 한다. 특히 판별분석의 성능과 연관이 있을 것으로 예상되는 레이블 노이즈의 발생 비율, 발생형태, 데이터의 개수에 따른 판별성능을 모의실험을 통해 살펴보았다. 그 결과, 데이터의 형태와 분석기법에 따라 레이블 노이즈가 판별성능에 영향을 미치는 정도가 다름을 확인하였다.

• 비파괴검사학회지
• /
• 제15권2호
• /
• pp.407-414
• /
• 1995

원자력발전소의 주요 기기들에 대한 가동중검사는 관련법규에 따라 철저히 수행되고 있다. 그러나 최근 선진국에서는 이에 만족하지 않고 원전 기기의 안전성을 더욱 확고히 하기 위해 기존의 가동중검사 요건을 계속 강화하고 있으며, 원전 관련 당사자들은 강화된 요건들을 충족시키기 위한 노력을 끊임없이 계속하고 있다. 이 글에서는 원전 기기 가동중검사 신규 요건들인 초음파탐상검사 시스템의 기량검증(Performance Demonstration) 요건, 비파괴검사자 및 초음파검사자 자격 인정 요건(ANSI/ASNT CP-189, Appendix VII of ASME Sec. XI), 증기발생기 전열관 와전류검사, 신호평가자 자격인정(Qualified Data Analyst : QDA), 미국규제기관(NRC)에서 발행하고 있는 NRC Bulletin, NRC information 등의 가동중검사 관련 사항들을 살펴보고 선진 외국에서는 이들 요건 및 정보에 대해 어떻게 대처하고 있는가를 알아본다. 또한 국내에서도 이들 신규 요건에 대한 대처 현황과 대처 방안을 모색한다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제8권1호
• /
• pp.1-5
• /
• 2008

Generally the analytical tools of data mining have two learning types which are supervised and unsupervised learning algorithms. Classification and prediction are main analysis tools for supervised learning. In this paper, we perform a comparison study of classification algorithms in data mining. We make comparative studies between popular classification algorithms which are LDA, QDA, kernel method, K-nearest neighbor, naive Bayesian, SVM, and CART. Also, we use almost all classification data sets of UCI machine learning repository for our experiments. According to our results, we are able to select proper algorithms for given classification data sets.

• 천문학논총
• /
• 제32권1호
• /
• pp.49-53
• /
• 2017

The point sources in the Bright Source Catalogue of the AKARI Far-Infrared Surveyor (FIS) were classified based on their FIR and mid-IR fluxes and colours into young stellar object (YSO) and extragalactic source types using a Quadratic Discriminant Analysis method (QDA) and Support Vector Machines (SVM). The reliability of the selection of YSO candidates is high, and the number of known YSO candidates were increased significantly, that we demonstrate in the case of the nearby open cluster IC348. Our results show that we can separate galactic and extragalactic AKARI point sources in the multidimensioal space of FIR fluxes and colours with high reliability, however, differentiating among the extragalactic sub-types needs further information.

• Communications for Statistical Applications and Methods
• /
• 제17권4호
• /
• pp.561-574
• /
• 2010

본 논문은 8가지 방법의 데이터 마이닝 알고리즘(CART, QUEST, CRUISE, 로지스틱 회귀분석, 선형판별분석, 이차판별분석, 신경망분석, 서포트 벡터 머신) 기법과 단일 알고리즘에 2가지 앙상블기법(배깅, 부스팅)을 적용한 16가지 방법을 바탕으로 총 24가지의 방법을 비교하였다. 알고리즘의 성능 비교를 위하여 13개의 이항반응변수로 구성된 데이터를 사용하였다. 비교 기준은 민감도, 특이도 및 오분류율을 사용하여 데이터 마이닝 기법의 성능향상에 대해 평가하였다.