• Korean Journal of Agricultural and Forest Meteorology
• /
• v.16 no.3
• /
• pp.157-168
• /
• 2014

From a complex systems perspective, ecohydrological systems in forests may be characterized with (1) large networks of components which give rise to complex collective behaviors, (2) sophisticated information processing, and (3) adaptation through self-organization and learning processes. In order to demonstrate such characteristics, we applied the recently proposed 'process networks' approach to a temperate deciduous forest in Gwangneung National Arboretum in Korea. The process network analysis clearly delineated the forest ecohydrological systems as the hierarchical networks of information flows and feedback loops with various time scales among different variables. Several subsystems were identified such as synoptic subsystem (SS), atmospheric boundary layer subsystem (ABLS), biophysical subsystem (BPS), and biophysicochemical subsystem (BPCS). These subsystems were assembled/disassembled through the couplings/decouplings of feedback loops to form/deform newly aggregated subsystems (e.g., regional subsystem) - an evidence for self-organizing processes of a complex system. Our results imply that, despite natural and human disturbances, ecosystems grow and develop through self-organization while maintaining dynamic equilibrium, thereby continuously adapting to environmental changes. Ecosystem integrity is preserved when the system's self-organizing processes are preserved, something that happens naturally if we maintain the context for self-organization. From this perspective, the process networks approach makes sense.

• Journal of the Korean Society for Railway
• /
• v.19 no.2
• /
• pp.145-158
• /
• 2016

This paper introduces a risk graph which is one method for determining the SIL as a measure of the effectiveness of signaling system. The purpose of this research is to make up for the weakness of the qualitative determination, which has input value ambiguity and a boundary problem in the SIL range. The fuzzy input valuable consists of consequence, exposure, avoidance and demand rate. The fuzzy inference produces forty eight fuzzy rule by adapting the calibrated risk graph in the IEC 61511. The Max-min composition is utilized for the fuzzy inference. The result of the fuzzy inference is the fuzzy value. Therefore, using the de-fuzzification method, the result should be converted to a crisp value that can be utilized for real projects. Ultimately, the safety requirement for hazard is identified by proposing a SIL result with a tolerable hazard rate. For the validation the results of the proposed method, the fuzzy risk graph model is compared with the safety analysis of the signaling system in CENELEC SC 9XA WG A10 report.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.2
• /
• pp.165-172
• /
• 2010

In clinical data minig, choosing the optimal subset of features is such important, not only to reduce the computational complexity but also to improve the usefulness of the model constructed from the given data. Moreover the threshold values (i.e., cut-off points) of selected features are used in a clinical decision criteria of experts for differential diagnosis of diseases. In this paper, we propose a fuzzy discretization approach, which is evaluated by measuring the degree of separation of redundant attribute values in overlapping region, based on spatial distribution of data with continuous attributes. The weighted average of the redundant attribute values is then used to determine the threshold value for each feature and rough set theory is utilized to select a subset of relevant features from the overall features. To verify the validity of the proposed method, we compared experimental results, which applied to classification problem using 668 patients with a chief complaint of dyspnea, based on three discretization methods (i.e., equal-width, equal-frequency, and entropy-based) and proposed discretization method. From the experimental results, we confirm that the discretization methods with fuzzy partition give better results in two evaluation measures, average classification accuracy and G-mean, than those with hard partition.

• The KIPS Transactions:PartB
• /
• v.12B no.7 s.103
• /
• pp.811-822
• /
• 2005

Automated text categorization is to classify free text documents into predefined categories automatically and whose main goals is to reduce considerable manual process required to the task. The researches to improving the text categorization performance(efficiency) in recent years, focused on enhancing existing classification models and algorithms itself, but, whose range had been limited by feature based statistical methodology. In this paper, we propose RTPost system of different style from i.ny traditional method, which takes fault tolerant system approach and data mining strategy. The 2 important parts of RTPost system are reinforcement training and post-processing part. First, the main point of training method deals with the problem of defining category to be classified before selecting training sample documents. And post-processing method deals with the problem of assigning category, not performance of classification algorithms. In experiments, we applied our system to documents getting low classification accuracy which were laid on a decision boundary nearby. Through the experiments, we shows that our system has high accuracy and stability in actual conditions. It wholly did not depend on some variables which are important influence to classification power such as number of training documents, selection problem and performance of classification algorithms. In addition, we can expect self learning effect which decrease the training cost and increase the training power with employing active learning advantage.

• Ecology and Resilient Infrastructure
• /
• v.2 no.3
• /
• pp.191-197
• /
• 2015

Invasive alien species are considered to be one of the main factors that cause biodiversity loss. Establishment of management strategies through continuous monitoring and risk assessment is a key element for invasive alien species management policy. In the present study, we introduce examples of ecological risk assessment tools developed in Japan, Germany-Austria and Belgium. Invasive alien species have been designated in Japan based on the assessment of risks to ecosystems, human health and primary industry. German-Austrian Black List Information System categorized alien species into Black List, White List and Grey List according to their risks to biodiversity. In the Harmonia Information System developed in Belgium, invasiveness, adverse impacts on native species and ecosystem functions and invasion stages were assessed and alien species were categorized into Black List, Watch List and Alert List. These international risk assessment tools may be helpful to improve our national risk assessment protocol for the prioritization of invasive alien species management.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.12
• /
• pp.1573-1579
• /
• 2013

To test the performance of nuclear fuels, coolant needs to be circulated through the test rig installed in the test loop. Because the pressure and temperature of the coolant is 15.5 MPa and $300^{\circ}C$ respectively, coolant sealing is one of the most important processes in fabricating a nuclear fuel test rig. In particular, 15 instrumentation cables installed in a test rig pass through the pressure boundary, and brazing is generally applied as a sealing method. In this study, an induction brazing system has been developed using a high frequency induction heater including a vacuum chamber. For application in the nuclear field, BNi2 should be used as a paste, and optimal process variables for Ni brazing have been found by several case studies. The performance and soundness of the brazed components has been verified by a tensile test, cross section test, and sealing performance test.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.5
• /
• pp.599-604
• /
• 2005

This paper presents fuzzy rules to predict diagnosis of Wisconsin breast cancer with minimized number of feature in put using the neural network with weighted fuzzy membership functions (NEWFM) and the non-overlap area distribution measurement method. NEWFM is capable of self-adapting weighted membership functions from the given the Wisconsin breast cancer clinical training data. n set of small, medium, and large weighted triangular membership functions in a hyperbox are used for representing n set of featured input. The membership functions are randomly distributed and weighted initially, and then their positions and weights are adjusted during learning. After learning, prediction rules are extracted directly from n set of enhanced bounded sums of n set of small, medium, and large weighted fuzzy membership functions. Then, the non-overlap area distribution measurement method is applied to select important features by deleting less important features. Two sets of prediction rules extracted from NEWFM using the selected 4 input features out of 9 features outperform to the current published results in number of set of rules, number of input features, and accuracy with 99.71%.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.416-416
• /
• 2019

육상 담수는 대기중 이산화탄소($CO_2$) 배출의 중요한 발생원으로 주목되고 있다. 하천 및 강에서 대기중으로 배출되는 $CO_2$는 전 세계 탄소순환의 핵심요소이며, 대부분의 하천과 강은 $CO_2$로 과포화 되어있다. 세계적으로 하천 및 강의 $CO_2$ 배출량은 호수 및 저수지의 배출량보다 약 5배 많은 것으로 보고되고 있으나, 국내연구에서는 연구사례가 드물다. 따라서 본 연구의 목적은 낙동강 중하류에 위치해있는 강정고령보(GGW), 달성보(DSW), 합천창녕보(HCW), 창녕함안보(CHW)에서 발생되는 순 대기 배출 플럭스(Net Atmospheric Flux, NAF)의 동적 변동 특성을 분석하고, 데이터마이닝 기법을 적용하여 쉽게 수집할 수 있는 물리적 및 수질 변수로 $CO_2$ NAF를 추정하는데 사용할 수 있는 간략한 예측 모델을 개발하는데 있다. $CO_2$ NAF는 대기-수면 경계면에서의 $CO_2$ 부분압($pCO_2$)의 차에 기체전달속도를 곱하여 산정하였으며, 기체전달속도는 Cole and Caraco(1998)가 제안한 식을 사용하였다. 담수와 해수의 탄산염 시스템에서 열역학적 화학평형을 모두 고려한 $CO_2$SYS 프로그램을 사용하여 수중의 $pCO_2$를 산정하였고, $CO_2$ NAF는 Henry의 법칙과 Fick의 1차 확산법칙을 사용하여 계산하였다. $CO_2$ NAF의 시간적 변동성에 영향을 미치는 환경요인을 평가하기 위해서 상관분석, 주성분분석(Principal Component Analysis; PCA), 단계적다중회귀모델(Step-wise Multiple Linear Regression; SMLR), 랜덤포레스트(Random Forest; RF)방법을 사용하였다. SMLR 모델은 R package인 olsrr, RF 모델은 R package인 caret, randomForest를 이용하여 분석하였다. 연구 결과, 4개 보 상류 하천구간은 조류의 성장이 활발한 일부 기간을 제외한 대부분의 기간에서 $CO_2$를 대기로 배출하는 종속영양시스템(Heterotrophic system)을 보였다. $CO_2$ NAF의 중위값은 HCW에서 최소 $391.5mg-CO_2/m^2day$, DSW에서 최대 $1472.7mg-CO_2/m^2day$였다. 모든 보에서 NAF는 pH와 강한 음의 상관관계를 보였으며, $pCO_2$와 Chl-a도 음의 상관관계를 보였다. 이는 조류가 수중에서 $CO_2$를 소비하고 pH를 증가시키기 때문이다. PCA 분석 결과, NAF와 $pCO_2$가 높은 공분산을 보였으며, pH와 Chl-a는 반대 방향으로 군집되어 상관분석과 동일한 결과를 보였다. 이 연구를 통해 개발된 SMLR 모델과 RF 모델의 Adj. $R^2$ 값은 모든 보에서 0.77 이상으로 나왔으며, $pCO_2$ 측정 데이터가 없더라도 하천의 $CO_2$ NAF를 추정하는 방법으로 사용될 수 있을 것으로 평가된다.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.11
• /
• pp.441-448
• /
• 2019

Previous researches on automatic spacing of Korean sentences has been researched to correct spacing errors by using n-gram based statistical techniques or morpheme analyzer to insert blanks in the word boundary. In this paper, we propose an end-to-end automatic word spacing by using deep neural network. Automatic word spacing problem could be defined as a tag classification problem in unit of syllable other than word. For contextual representation between syllables, Bi-LSTM encodes the dependency relationship between syllables into a fixed-length vector of continuous vector space using forward and backward LSTM cell. In order to conduct automatic word spacing of Korean sentences, after a fixed-length contextual vector by Bi-LSTM is classified into auto-spacing tag(B or I), the blank is inserted in the front of B tag. For tag classification method, we compose three types of classification neural networks. One is feedforward neural network, another is neural network language model and the other is linear-chain CRF. To compare our models, we measure the performance of automatic word spacing depending on the three of classification networks. linear-chain CRF of them used as classification neural network shows better performance than other models. We used KCC150 corpus as a training and testing data.

• Korean Journal of Remote Sensing
• /
• v.37 no.5_2
• /
• pp.1269-1279
• /
• 2021

Geostationary Ocean Color Imager II (GOCI-II), which are now operated successfully since its launch in 2020, acquires local area images with 12 Level 1B slot images that are sequentially acquired in a 3×4 grid pattern. The boundary areas between the adjacent slots are prone to discontinuity in radiance, which becomes even more clear in the following Level 2 data, and this warrants the precise analysis and correction before the distribution. This study evaluates the relative radiometric biases between the adjacent slots images, by exploiting the overlapped areas across the images. Although it is ideal to derive the statistics from humongous images, this preliminary analysis uses just the scenes acquired at a specific time to understand its general behavior in terms of bias and variance in radiance. Level 1B images of February 21st, 2021 (UTC03 = noon in local time) were selected for the analysis based on the cloud cover, and the radiance statistics were calculated only with the ocean pixels. The results showed that the relative bias is 0~1% in all bands but Band 1 (380 nm), while Band 1 exhibited a larger bias (1~2%). Except for the Band 1 in slot pairs aligned North-South, biases in all direction and in all bands turned out to have biases in the opposite direction that the sun elevation would have caused.