• 제목/요약/키워드: similarity dimension

검색결과 133건 처리시간 0.025초

새로운 시간축 정규화 방법을 이용한 한국어 고립단어 인식기 (Korean isolated word recognizer using new time alignment method of speech signal)

  • 남명우;박규홍;노승용
    • 대한전자공학회논문지SP
    • /
    • 제38권5호
    • /
    • pp.567-575
    • /
    • 2001
  • 본 논문에서는 음성신호의 발성길이와 상관없이 일정한 크기의 파라미터를 얻을 수 있는 새로운 방법을 제안하였다. 음성인식기의 성능은 음성신호에서 추출된 파라미터간의 유사도(패턴간의 거리)를 어떻게 비교하는지에 따라 결정된다. 그러나 화자에 따른 음성신호의 변이나 발성속도의 차이는 음성신호에서 일정한 크기의 파라미터 추출을 어렵게 한다. 제안한 방법은 음성신호에서 얻어진 파라미터를 스펙토그램의 형태로 표현한 뒤 2차원 DCT(Discrete Cosine Transform)를 이용해 일정한 크기의 파라미터로 정규화시키는 방법이다. 제안한 방법의 유효성을 입증하기 위해 청각세포를 모델링한 32개의 대역통과 필터로부터 얻어진 음성신호의 파라미터를 2차원 DCT 방법으로 가공한 후, 신경 회로망의 입력으로 사용하였다. 또한 기존 방법과의 인식률 비교를 위해 기존의 정규화된 입력을 구하는 방법 중 하나를 선택하여 비교 실험을 수행하였다. 실험결과 제안한 방법은 기존 방법에 비해 화자종속 및 화자독립 고립단어 인식에서 더 높은 인식률과 빠른 인식속도를 얻을 수 있었다.

  • PDF

Vertical and longitudinal variations in plant communities of drawdown zone of a monsoonal riverine reservoir in South Korea

  • Cho, Hyunsuk;Marrs, Rob H.;Alday, Josu G.;Cho, Kang-Hyun
    • Journal of Ecology and Environment
    • /
    • 제43권2호
    • /
    • pp.271-281
    • /
    • 2019
  • Background: The plant communities within reservoir drawdown zones are ecologically important as they provide a range of ecosystem services such as stabilizing the shoreline, improving water quality, enhancing biodiversity, and mitigating climate change. The aim of the study was therefore to identify the major environmental factors affecting these plant communities within the drawdown zone of the Soyangho Reservoir in South Korea, which experiences a monsoonal climate, and thereafter to (1) elucidate the plant species responses and (2) compare the soil seedbank composition along main environmental gradients. Results: Two main environmental gradients affecting the plant community structure were identified within the drawdown zone; these were a vertical and longitudinal gradient. On the vertical dimension, a hydrological gradient of flood/exposure, the annual-dominated plant community near the water edge changed to a perennial-dominated community at the highest elevation. On the longitudinal dimension from the dam to the upstream, plant species composition changed from an upland forest-edge community to a lowland riverine community, and this was correlated with slope degree, soil particle size, and soil moisture content. Simultaneously, the composition of the soil seedbank was separated along the vertical gradient of the drawdown zone, with mainly annuals near the water edge and some perennials at higher elevations. The species composition similarity between the seedbank and extant vegetation was greater in the annual communities at low elevation than in the perennial communities at higher elevation. Conclusions: The structures of plant community and soil seedbank in the drawdown zone of a monsoonal riverine reservoir were changed first along the vertical and secondly along the longitudinal gradients. The soil seedbank could play an important role on the vegetation regeneration after the disturbances of flood/exposure in the drawdown zone. These results indicate that it is important to understand the vertical and longitudinal environmental gradients affecting shoreline plant community structure and the role of soil seedbanks on the rapid vegetation regeneration for conserving and restoring the drawdown zone of a monsoonal reservoir.

SNS대상의 지능형 자연어 수집, 처리 시스템 구현을 통한 한국형 감성사전 구축에 관한 연구 (Research on Designing Korean Emotional Dictionary using Intelligent Natural Language Crawling System in SNS)

  • 이종화
    • 한국정보시스템학회지:정보시스템연구
    • /
    • 제29권3호
    • /
    • pp.237-251
    • /
    • 2020
  • Purpose The research was studied the hierarchical Hangul emotion index by organizing all the emotions which SNS users are thinking. As a preliminary study by the researcher, the English-based Plutchick (1980)'s emotional standard was reinterpreted in Korean, and a hashtag with implicit meaning on SNS was studied. To build a multidimensional emotion dictionary and classify three-dimensional emotions, an emotion seed was selected for the composition of seven emotion sets, and an emotion word dictionary was constructed by collecting SNS hashtags derived from each emotion seed. We also want to explore the priority of each Hangul emotion index. Design/methodology/approach In the process of transforming the matrix through the vector process of words constituting the sentence, weights were extracted using TF-IDF (Term Frequency Inverse Document Frequency), and the dimension reduction technique of the matrix in the emotion set was NMF (Nonnegative Matrix Factorization) algorithm. The emotional dimension was solved by using the characteristic value of the emotional word. The cosine distance algorithm was used to measure the distance between vectors by measuring the similarity of emotion words in the emotion set. Findings Customer needs analysis is a force to read changes in emotions, and Korean emotion word research is the customer's needs. In addition, the ranking of the emotion words within the emotion set will be a special criterion for reading the depth of the emotion. The sentiment index study of this research believes that by providing companies with effective information for emotional marketing, new business opportunities will be expanded and valued. In addition, if the emotion dictionary is eventually connected to the emotional DNA of the product, it will be possible to define the "emotional DNA", which is a set of emotions that the product should have.

The first insight into the structure of the Photosystem II reaction centre complex at $6{\AA}$ resolution determined by electron crystallography

  • Rhee, Kyong-Hi
    • 한국식물학회:학술대회논문집
    • /
    • 한국식물학회 1999년도 Proceedings of the 17th Symposium on Plant Biology Environmental Stress and Photosynthesis
    • /
    • pp.83-90
    • /
    • 1999
  • Electron crystallography of two-dimensional crystalsand electron cryo-microscopy is becoming an established method for determining the structure and function of a variety of membrane proteins that are providing difficult to crystallize in three dimension. In this study this technique has been used to investigate the structure of a ~160 kDa reaction centre sub-core complex of photosystem II. Photosystem II is a photosynthetic membrane protein consisting of more than 25 subunits. It uses solar energy to split water releasing molecular oxygen into the atmosphere and creates electrochemical potential across the thylakoid membrane, which is eventually utilized to generate ATP and NADPH. Images were taken using Philips CM200 field emission gun electron microscope with an acceleration voltage of 200kW at liquid nitrogen temperature. In total, 79 images recorded dat tilt angles ranging from 0 to 67 degree yielded amplitudes and phases for a three-dimensional map with an in-plant resolution of 6$\AA$ and 11.4$\AA$ in the third dimension shows at least 23 transmembrane helices resolved in a monomeric complex, of which 18 were able to be assigned to the D1, D2, CP47 , and cytochrome b559 alfa beta-subunits with their associated pigments that ae active in electron transport (Rhee, 1998, Ph.D.thesis). The D1/D2 heterodimer is located in the central position within the complex and its helical scalffold is remarkably similar to that of the reaction centres not only in purple bacteria but also in plant photosystem I (PSI) , indicating a common evoluationary origin of all types of reaction centre in photosynthetic organism known today 9RHee et al. 1998). The structural homology is now extended to the inner antenna subunit, ascribed to CP47 in our map, where the 6 transmembrane helices show a striking structural similarity to the corresponding helices of the PSI reaction centre proteins. The overall arrangement of the chlorophylls in the D1 /D2 heterodimer, and in particular the distance between the central pair, is ocnsistent with the weak exciton coupling of P680 that distinguishes this reaction centre from bacterial counterpart. The map in most progress towards high resolution structure will be presented and discussed.

  • PDF

문장 분류를 위한 정보 이득 및 유사도에 따른 단어 제거와 선택적 단어 임베딩 방안 (Selective Word Embedding for Sentence Classification by Considering Information Gain and Word Similarity)

  • 이민석;양석우;이홍주
    • 지능정보연구
    • /
    • 제25권4호
    • /
    • pp.105-122
    • /
    • 2019
  • 텍스트 데이터가 특정 범주에 속하는지 판별하는 문장 분류에서, 문장의 특징을 어떻게 표현하고 어떤 특징을 선택할 것인가는 분류기의 성능에 많은 영향을 미친다. 특징 선택의 목적은 차원을 축소하여도 데이터를 잘 설명할 수 있는 방안을 찾아내는 것이다. 다양한 방법이 제시되어 왔으며 Fisher Score나 정보 이득(Information Gain) 알고리즘 등을 통해 특징을 선택 하거나 문맥의 의미와 통사론적 정보를 가지는 Word2Vec 모델로 학습된 단어들을 벡터로 표현하여 차원을 축소하는 방안이 활발하게 연구되었다. 사전에 정의된 단어의 긍정 및 부정 점수에 따라 단어의 임베딩을 수정하는 방법 또한 시도하였다. 본 연구는 문장 분류 문제에 대해 선택적 단어 제거를 수행하고 임베딩을 적용하여 문장 분류 정확도를 향상시키는 방안을 제안한다. 텍스트 데이터에서 정보 이득 값이 낮은 단어들을 제거하고 단어 임베딩을 적용하는 방식과, 정보이득 값이 낮은 단어와 코사인 유사도가 높은 주변 단어를 추가로 선택하여 텍스트 데이터에서 제거하고 단어 임베딩을 재구성하는 방식이다. 본 연구에서 제안하는 방안을 수행함에 있어 데이터는 Amazon.com의 'Kindle' 제품에 대한 고객리뷰, IMDB의 영화리뷰, Yelp의 사용자 리뷰를 사용하였다. Amazon.com의 리뷰 데이터는 유용한 득표수가 5개 이상을 만족하고, 전체 득표 중 유용한 득표의 비율이 70% 이상인 리뷰에 대해 유용한 리뷰라고 판단하였다. Yelp의 경우는 유용한 득표수가 5개 이상인 리뷰 약 75만개 중 10만개를 무작위 추출하였다. 학습에 사용한 딥러닝 모델은 CNN, Attention-Based Bidirectional LSTM을 사용하였고, 단어 임베딩은 Word2Vec과 GloVe를 사용하였다. 단어 제거를 수행하지 않고 Word2Vec 및 GloVe 임베딩을 적용한 경우와 본 연구에서 제안하는 선택적으로 단어 제거를 수행하고 Word2Vec 임베딩을 적용한 경우를 비교하여 통계적 유의성을 검정하였다.

펄프·제지 산업에서의 프랙탈 기하 원리 및 그 응용 (The Principles of Fractal Geometry and Its Applications for Pulp & Paper Industry)

  • 고영찬;박종문;신수정
    • 펄프종이기술
    • /
    • 제47권4호
    • /
    • pp.177-186
    • /
    • 2015
  • Until Mandelbrot introduced the concept of fractal geometry and fractal dimension in early 1970s, it has been generally considered that the geometry of nature should be too complex and irregular to describe analytically or mathematically. Here fractal dimension indicates a non-integer number such as 0.5, 1.5, or 2.5 instead of only integers used in the traditional Euclidean geometry, i.e., 0 for point, 1 for line, 2 for area, and 3 for volume. Since his pioneering work on fractal geometry, the geometry of nature has been found fractal. Mandelbrot introduced the concept of fractal geometry. For example, fractal geometry has been found in mountains, coastlines, clouds, lightning, earthquakes, turbulence, trees and plants. Even human organs are found to be fractal. This suggests that the fractal geometry should be the law for Nature rather than the exception. Fractal geometry has a hierarchical structure consisting of the elements having the same shape, but the different sizes from the largest to the smallest. Thus, fractal geometry can be characterized by the similarity and hierarchical structure. A process requires driving energy to proceed. Otherwise, the process would stop. A hierarchical structure is considered ideal to generate such driving force. This explains why natural process or phenomena such as lightning, thunderstorm, earth quakes, and turbulence has fractal geometry. It would not be surprising to find that even the human organs such as the brain, the lung, and the circulatory system have fractal geometry. Until now, a normal frequency distribution (or Gaussian frequency distribution) has been commonly used to describe frequencies of an object. However, a log-normal frequency distribution has been most frequently found in natural phenomena and chemical processes such as corrosion and coagulation. It can be mathematically shown that if an object has a log-normal frequency distribution, it has fractal geometry. In other words, these two go hand in hand. Lastly, applying fractal principles is discussed, focusing on pulp and paper industry. The principles should be applicable to characterizing surface roughness, particle size distributions, and formation. They should be also applicable to wet-end chemistry for ideal mixing, felt and fabric design for papermaking process, dewatering, drying, creping, and post-converting such as laminating, embossing, and printing.

지수가중 이동평균 기반의 PPG 신호 동잡음 제거 (The Motion Artifact Reduction from the PPG based on EWMA)

  • 이준연
    • 디지털융복합연구
    • /
    • 제11권8호
    • /
    • pp.183-190
    • /
    • 2013
  • PPG 신호는 심장의 박동에 동기된 유사 주기 신호이다. 본 논문에서는 PPG 신호의 유사주기성을 이용한 지수가중 이동평균필터 방법을 제안한다. 이 필터링 방법은 PPG 신호를 주기적으로 분리하여 각 주기 신호의 같은 순번에 있는 샘플들끼리 평균을 취하는 방법이다. 연속된 PPG 신호의 주기중에 동잡음이 혼입되었다면 주기를 기준으로 PPG 신호를 분리한 후, 각 주기의 샘플수를 조정하여 같은 샘플수를 가지게 만든다. 이 주기들을 2차원으로 배열한 후 현재 주기부터 이전 각 주기의 샘플끼리 평균을 취함으로써 훼손없이 동잡음을 제거할 수 있었다.

한글 감정단어의 의미적 관계와 범주 분석에 관한 연구 (A Study on the Analysis of Semantic Relation and Category of the Korean Emotion Words)

  • 이수상
    • 한국도서관정보학회지
    • /
    • 제47권2호
    • /
    • pp.51-70
    • /
    • 2016
  • 이 연구의 목적은 한글로 된 주요감정단어들의 리스트를 대상으로 의미적 관계의 네트워크와 극성과 각성의 범주를 분석하는데 있다. 분석결과는 다음과 같다. 첫째, 감정단어 네트워크에서 각 감정단어들은 의미적으로 연결되어 있었다. 이것은 의미적 유사성에 따라 감정단어들의 유형을 구분하는 것을 어렵게 하는 특징이다. 대신에 의미적 관계의 감정단어 네트워크에서 중심적인 역할을 수행하는 감정단어들을 확인할 수 있었다. 둘째, 극성과 각성의 차원을 혼합한 범주에서, 많은 감정단어들은 부정적인 극성과 높은 각성의 단어들 집단과 부정적인 극성과 중간수준 각성의 단어들 집단으로 분류되었다. 이러한 한글감정단어의 특성들은 도서관이나 문헌정보에 나타나는 각종 텍스트 데이터의 감정분석에 유용하게 활용될 것이다.

한국 제조업의 지식 네트워크의 구조적 변화의 특성 (The Characteristics of Structural Charge in Knowledge Network of Korean Manufacturing)

  • 김문수;오형식;박용태
    • 기술경영경제학회:학술대회논문집
    • /
    • 기술경영경제학회 1997년도 제12회 동계학술발표회 논문집
    • /
    • pp.133-158
    • /
    • 1997
  • This paper analyzes the characteristics of technological knowledge flow-structure of Korean manufacturing in dynamic perspective. In doing that, the concept of the knowledge network is introduced which is defined as a set of industries and their interaction(knowledge flow) or linkage. The analysis of the inter-industrial knowledge flows is based on the technological similarity by using R&D researchers'academic background in the year of 1984, 1987, 1990. The analysis is carried out by such methodology as network analysis, indicator analysis and simple statistical analysis. And the final results are drawn both in absolute terms(dimension effect) and in relative terms (proportion effect) respectively. The main findings are as follow. First, the Korean manufacturing knowledge network appears to strengthen existing inter-industrial knowledge linkages rather than to construct new linkages. Second, the network seems to form a dualistic structure in that some high-technology sectors (knowledge production sectors) emerge along with traditional sectors (knowledge absorbing sectors). Third, since the mid-1980s, an inter-industrial fusion is witnessed among technologically intensive sectors, indicating that some sophisticated innovation modes are emerging in Korean manufacturing system.

  • PDF

한국 제조업 지식네트워크 구조변화의 특성 (The Characteristics of Structural Change in Knowledge Network of Korean Manufacturing Industries)

  • 김문수;오형식;박용태
    • 기술혁신연구
    • /
    • 제6권1호
    • /
    • pp.71-98
    • /
    • 1998
  • This paper analyzes the characteristics of technological knowledge flow-structure of Korean manufacturing in dynamic perspective. In doing that, the concept of the knowledge network is introduced which is defined as a set of industries and their interaction(knowledge flow) or linkage. The analysis of the inter-industrial knowledge flows is based on the technological similarity by using R&D researchers' academic background in the year of 1984, 1987, 1990. The analysis is carried out by such methodology as network analysis, indicator analysis and simple statistical analysis. And the final results are drawn both in absolute terms(dimension effect) and in relative terms(proportion effect) respectively. The main findings are as follow. First, the Korean manufacturing knowledge network appears to strengthen existing inter-industrial knowledge linkages rather than to construct new linkages. Second, the network seems to form a dualistic structure in that some high-technology sectors(knowledge production sectors) emerge along with traditional sectors(knowledge absorbing sectors). Third, since the mid-1980s, an inter-industrial fusion is witnessed among technologically intensive sectors, indicating that some sophisticated innovation modes are emerging in Korean manufacturing system. And fourth, by using the relations of the inter-industrial knowledge-flows, we classified manufacturing industries into 3 type ; knowledge-outflow sector, knowledge-inflow sector and knowledge intermediary sector.

  • PDF