• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1032-1032
• /
• 2001

Artificial Neural Network (ANN) calibration techniques have been used commercially for agricultural applications since the mid-nineties. Global models, based on transmission data from 850 to 1050 nm, are used routinely to measure protein and moisture in wheat and barley and also moisture in triticale, rye, and oats. These models are currently used commercially in approx. 15 countries throughout the world. Results concerning earlier European ANN models are being published elsewhere. Some of the findings from that study will be discussed here. ANN models have also been developed for coarsely ground samples of compound feed and feed ingredients, again measured in transmission mode from 850 to 1050 nm. The performance of models for pig- and poultry feed will be discussed briefly. These models were developed from a very large data set (more than 20,000 records), and cover a very broad range of finished products. The prediction curves are linear over the entire range for protein, fat moisture, fibre, and starch (measured only on poultry feed), and accuracy is in line with the performance of smaller models based on Partial Least Squares (PLS). A simple bias adjustment is sufficient for calibration transfer across instruments. Recently, we have investigated the possible use of ANN for a different type of NIR spectrometer, based on reflectance data from 1100 to 2500 nm. In one study, based on data for protein, fat, and moisture measured on unground compound feed samples, dedicated ANN models for specific product classes (cattle feed, pig feed, broiler feed, and layers feed) gave moderately better Standard Errors of Prediction (SEP) compared to modified PLS (MPLS). However, if the four product classes were combined into one general calibration model, the performance of the ANN model deteriorated only slightly compared to the class-specific models, while the SEP values for the MPLS predictions doubled. Brix value in molasses is a measure of sugar content. Even with a huge dataset, PLS models were not sufficiently accurate for commercial use. In contrast an ANN model based on the same data improved the accuracy considerably and straightened out non-linearity in the prediction plot. The work of Mr. David Funk (GIPSA, U. S. Department of Agriculture) who has studied the influence of various types of spectral distortions on ANN- and PLS models, thereby providing comparative information on the robustness of these models towards instrument differences, will be discussed. This study was based on data from different classes of North American wheat measured in transmission from 850 to 1050 nm. The distortions studied included the effect of absorbance offset pathlength variation, presence of stray light bandwidth, and wavelength stretch and offset (either individually or combined). It was shown that a global ANN model was much less sensitive to most perturbations than class-specific GIPSA PLS calibrations. It is concluded that ANN models based on large data sets offer substantial advantages over PLS models with respect to accuracy, range of materials that can be handled by a single calibration, stability, transferability, and sensitivity to perturbations.

• Journal of KIISE
• /
• v.44 no.6
• /
• pp.587-594
• /
• 2017

Recently, motor imagery electroencephalogram(EEG) based Brain-Computer Interface(BCI) systems have received a significant amount of attention in various fields, including medicine and engineering. The Common Spatial Pattern(CSP) algorithm is the most commonly-used method to extract the features from motor imagery EEG. However, the CSP algorithm has limited applicability in Small-Sample Setting(SSS) situations because these situations rely on a covariance matrix. In addition, large differences in performance depend on the frequency bands that are being used. To address these problems, 4-40Hz band EEG signals are divided using nine filter-banks and Regularized CSP(R-CSP) is applied to individual frequency bands. Then, the Mutual Information-Based Individual Feature(MIBIF) algorithm is applied to the features of R-CSP for selecting discriminative features. Thereafter, selected features are used as inputs of the classifier Least Square Support Vector Machine(LS-SVM). The proposed method yielded a classification accuracy of 87.5%, 100%, 63.78%, 82.14%, and 86.11% in five subjects("aa", "al", "av", "aw", and "ay", respectively) for BCI competition III dataset IVa by using 18 channels in the vicinity of the motor area of the cerebral cortex. The proposed method improved the mean classification accuracy by 16.21%, 10.77% and 3.32% compared to the CSP, R-CSP and FBCSP, respectively The proposed method shows a particularly excellent performance in the SSS situation.

• Journal of Navigation and Port Research
• /
• v.36 no.6
• /
• pp.521-526
• /
• 2012

More efficient and effective volume management of trade cargo is recently requested due to FTA with foreign country. Above all, the grain is the main cargo needed in Korean food life and was appointed as the core trade cargo during FTA. This study is aimed to forecast future demands of grain volumes which are handled at Incheon port because most of the grain volumes are traded at Incheon port in Korea. System Dynamics (SD) was used for forecasting as the methodology. Also, population, yearly grain consumption per a man, GDP, GRDP, exchange rate, and BDI were used as the factors that influence grain volumes. Simulation duration was from 2000 to 2020 and real data was used from 2000 to 2007. According to the simulation, 2020's grain volumes at Incheon port were forecasted to be about 2 million tons and grain volumes handled at Incheon port were continuously reduced. In order to measure accuracy of the simulation, this study implemented MAPE analysis. And after the implementation, the simulation was decided as a much more accurate model because MAPE value was calculated to be 6.3%. This study respectively examined factors using the sensitivity analysis. As a result, in terms of the effects on grain volume in Incheon Port, the population factor is most significant and exchange rate factor is the least.

• The Journal of Korean Academy of Prosthodontics
• /
• v.45 no.4
• /
• pp.431-443
• /
• 2007

Statement of problem: Injuries along with discomfort may result on the oral mucosa when non-rigid material is used as the major connector in construction of RPD, since nonrigid major connectors transmit unstable forces throughout the appliance. Titanium which recently draws attention as a substitute of Co-Cr had a difficulty in fabricating due to high melting temperature but the development of casting technique makes it possible to apply to the clinical case. Purpose: The purpose of this study was to investigate the rigidity and the castability of titanium upper major connector by design and make a comparison with Co-Cr major connectors which are widely used in clinical cases now. Material and methods: Casting was done using CP-Ti(Grage 2) (Kobe still Co., Japan) for the experimental groups, and 4 various designs namely palatal strap, U-shaped bar, A-P strap, and complete palatal plate were casted and 5 of each designs were included in each group. For the experimental group, Universal testing machine (Model 4502; Instron, Canton, Mass) was used to apply vertical torsional force vertically to the horizontal plane of major connector. In the second experiment, Vertical compressive force was applied to the horizontal plane of major connector. As a comparative group, Co-Cr major connector was equally manufactured and underwent the same experimental procedures Strain rate was measured after constant loading for one minute duration, and statistical analysis was done with SPSS ver.10.0 for WIN(SPSS. Inc. USA). From the one-way ANOVA and variance analysis (P=0.05), Scheffe's multiple comparison test implemented. Results: 1. Least amount of strain was observed with complete palatal plate followed by A-P bar, palatal bar, and the U-shaped bar having most amount of strain. 2. In all designs of titanium major connector, less strain rate was observed under compressive loading than under torsional loading showing more resistance to lateral force. 3. For titanium major connector, less strain rate was observed when the force is applied to the first premolar area rather than to the second molar area indicating more strength with shorter length of lever. 4. In Comparison of Co-Cr major connector with titanium major connector, palatal strap and U-shaped bar designs showed higher strength under torsional force that is statically significant, and under compressive force, no significant difference was observed expert for U-shaped bar. 5. In titanium major connector, complete palatal plate showed lowest success rate in casting when compared with the Co-Cr major connector. Conclusion: Above results prove that when using titanium for major connector, only with designs capable of generating rigidity can the major connector have almost equal amount of rigidity as Co-Cr major connector and show lower success rate in casting when compared with the Co-Cr major connector.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.4
• /
• pp.391-397
• /
• 2017

Recently, digital image correlation (DIC) techniques have been used to measure dynamic deformation during tensile testing. The standard tensile test method measures the average displacement of the relevant specimen to calculate the true stress-strain curve. Therefore, the validity of the true stress curve is restricted to the stress incurred within the uniform stretching interval, i.e., the maximum stress corresponds to the starting point of the necking deformation. Alternatively, if DIC is used, the effective range of the strain and strain rate can be extended to the breaking point of the tensile specimen, because of the feasibility of measuring the local strain over the entire area of interest. Because of these advantages, many optical 3D measurement systems have been introduced and used in research and industry. However, the conventional 3D measurement systems are exceedingly expensive and time consuming. In addition, these systems have the disadvantage of a very large equipment size which makes their transport difficult. In this study, a 2D image correction method employing a 2D DIC measurement method in conjunction with a numerical analysis method is developed using a smartphone. The results of the proposed modified 2D DIC method yielded higher accuracy than that obtained via the 3D measurement equipment. In conclusion, it was demonstrated that the proposed 2D DIC and calibration methods yield accurate measurement results with low time costs.

• Journal of Biomedical Engineering Research
• /
• v.42 no.6
• /
• pp.259-267
• /
• 2021

Recently, the market for personal health care and medical devices based on Bluetooth Low Energy(BLE) has grown rapidly. BLE is being used in various medical data communication devices based on low power consumption and universal compatibility. However, since data errors occurring in the transmission of medical data can lead to medical accidents, it is necessary to analyze the causes of errors and study methods to reduce data error. In this paper, the minimum communication speed to be used in medical devices was set to at least 800 byte/sec based on the wireless electrocardiography regulations of the Ministry of Food and Drug Safety. And the data loss rate was tested when data was transmitted at a speed higher than 800 byte/sec. The factors that cause communication data error were classified, and the relationship between each factor and the data error rate was analyzed through experiments. When there were two or more activated peripherals connected to the central, data error occurred due to channel hopping and bottleneck, and the data error rate increased in proportion to the communication distance and the number of activated peripherals. Through this experiment, when the BLE is used in a medical device that intermittently transmits biosignal data, the risk of a medical accident is predicted to be low if the number of peripherals is 3 or less. But, it was determined that BLE would not be suitable for the development of a biosignal measuring device that must be continuously transmitted in real time, such as an electrocardiogram.

• Korean Chemical Engineering Research
• /
• v.54 no.5
• /
• pp.621-629
• /
• 2016

Recently, the researches on quantitative structure activity relationship (QSAR) for describing toxicities or activities of chemicals based on chemical structural characteristics have been widely carried out in order to estimate the toxicity of chemicals in multiuse facilities. Because the toxicity of chemicals are explained by various kinds of molecular descriptors, an important step for QSAR model development is how to select significant molecular descriptors. This research proposes a statistical selection of significant molecular descriptors and a new QSAR model based on partial least square (PLS). The proposed QSAR model is applied to estimate the logarithm of partition coefficients (log P) of 130 polychlorinated biphenyls (PCBs) and lethal concentration ($LC_{50}$) of 14 PCBs, where the prediction accuracies of the proposed QSAR model are compared to a conventional QSAR model provided by OECD QSAR toolbox. For the selection of significant molecular descriptors that have high correlation with molecular descriptors and activity information of the chemicals of interest, correlation coefficient (r) and variable importance of projection (VIP) are applied and then PLS model of the selected molecular descriptors and activity information is used to predict toxicities and activity information of chemicals. In the prediction results of coefficient of regression ($R^2$) and prediction residual error sum of square (PRESS), the proposed QSAR model showed improved prediction performances of log P and $LC_{50}$ by 26% and 91% than the conventional QSAR model, respectively. The proposed QSAR method based on computational toxicology can improve the prediction performance of the toxicities and the activity information of chemicals, which can contribute to the health and environmental risk assessment of toxic chemicals.

• Journal of Intelligence and Information Systems
• /
• v.21 no.3
• /
• pp.101-116
• /
• 2015

Recently, due to the development of smart devices and social media, vast amounts of information with the various forms were accumulated. Particularly, considerable research efforts are being directed towards analyzing unstructured big data to resolve various social problems. Accordingly, focus of data-driven decision-making is being moved from structured data analysis to unstructured one. Also, in the field of recommendation system, which is the typical area of data-driven decision-making, the need of using unstructured data has been steadily increased to improve system performance. Approaches to improve the performance of recommendation systems can be found in two aspects- improving algorithms and acquiring useful data with high quality. Traditionally, most efforts to improve the performance of recommendation system were made by the former approach, while the latter approach has not attracted much attention relatively. In this sense, efforts to utilize unstructured data from variable sources are very timely and necessary. Particularly, as the interests of users are directly connected with their needs, identifying the interests of the user through unstructured big data analysis can be a crew for improving performance of recommendation systems. In this sense, this study proposes the methodology of improving recommendation system by measuring interests of the user. Specially, this study proposes the method to quantify interests of the user by analyzing user's internet usage patterns, and to predict user's repurchase based upon the discovered preferences. There are two important modules in this study. The first module predicts repurchase probability of each category through analyzing users' purchase history. We include the first module to our research scope for comparing the accuracy of traditional purchase-based prediction model to our new model presented in the second module. This procedure extracts purchase history of users. The core part of our methodology is in the second module. This module extracts users' interests by analyzing news articles the users have read. The second module constructs a correspondence matrix between topics and news articles by performing topic modeling on real world news articles. And then, the module analyzes users' news access patterns and then constructs a correspondence matrix between articles and users. After that, by merging the results of the previous processes in the second module, we can obtain a correspondence matrix between users and topics. This matrix describes users' interests in a structured manner. Finally, by using the matrix, the second module builds a model for predicting repurchase probability of each category. In this paper, we also provide experimental results of our performance evaluation. The outline of data used our experiments is as follows. We acquired web transaction data of 5,000 panels from a company that is specialized to analyzing ranks of internet sites. At first we extracted 15,000 URLs of news articles published from July 2012 to June 2013 from the original data and we crawled main contents of the news articles. After that we selected 2,615 users who have read at least one of the extracted news articles. Among the 2,615 users, we discovered that the number of target users who purchase at least one items from our target shopping mall 'G' is 359. In the experiments, we analyzed purchase history and news access records of the 359 internet users. From the performance evaluation, we found that our prediction model using both users' interests and purchase history outperforms a prediction model using only users' purchase history from a view point of misclassification ratio. In detail, our model outperformed the traditional one in appliance, beauty, computer, culture, digital, fashion, and sports categories when artificial neural network based models were used. Similarly, our model outperformed the traditional one in beauty, computer, digital, fashion, food, and furniture categories when decision tree based models were used although the improvement is very small.

• The Journal of Fisheries Business Administration
• /
• v.44 no.3
• /
• pp.15-28
• /
• 2013

Abalone aquaculture has developed very rapidly in Korea. Until the mid 1990s it has annually produced about 100 tons. Since then the yield has increased to about 9,000 tons in 2012. The amount accounts for 20% of the global abalone yield. About 86% of produced abalone is consumed domestically and the rest is exported. 100 tons for export seemed as an unattainable goal back in 2003. However, the export rose up to 1,333 tons in 2012. Despite its rapid growth, Korean abalone industry is faced with some problems. The first is the slowdown of yield increase rates. Abalone production increased by 50~60% until the mid 2000. However, the rate continued to drop to below 10%. Reasons behind the slow increase are deteriorating aquaculture grounds and worsening market problems. Constant aquaculture aggravated productivity and overcrowded facilities at a limited space made matters worse. Moreover, abalone export has stalled and so did domestic consumption. In the meantime, rising mortality of young abalone has lowered productivity at abalone breeding places. The mortality rates of abalone remained below 5% in the early 2000s but rose to 30~40% these days. This translates into rising abalone prices. The market problems imply stagnant or shrinking export as well as domestic consumption. The export increase rates took a nosedive from 200 to below 50 between the early 2000s and the late 2000s. Moreover, the increase rates of domestic consumption have become remarkably sluggish. According to, it stood at 50~60% in the mid 2000s but continued to decrease after 2008. These problems, in turn, affected the size of abalone. The usual abalone size for market was 10~12 shells per kg, but recently the size became smaller and smaller to 15~16 shells per kg. The change of size implies shift in consumption patterns: Consumers not only eat live abalone but also they cook soup with it. The size of abalone for uncooked dish is usually very big, like 10~12 shells per kg. In contrast, smaller abalone, such as 20~25 shells per kg, are used for making soup. Increasing use of smaller abalone leads to lower income of abalone aquaculture households. This is partly because that the size determines the price and the price gap between big abalones and smaller ones is extreme in Korea. For the sustainable growth of Korean abalone industry, we need to come up with strategies. First, a reasonable production system needs to be in place, especially for better management of abalone aquaculture grounds. Management of abalone licenses is also necessary because local governments issue relevant licenses as well as supervising abalone grounds. Second, abalone export destination need to be diversified. Japan, the major importer of Korean abalone, takes up a lion's share of export, at 95%. Third, new consumption style of abalone needs to be developed. Abalone used to be consumed as 'raw type' or Sashimi in Korea. This sole type of consumption hampers the growth of abalone market. Moreover, more strategies are needed to encourage and distribute home cooking of abalone rather than eating-out at restaurants. Last but not least, distribution system should be improved for better delivery of live abalone.

• Korean Journal of Clinical Pharmacy
• /
• v.23 no.4
• /
• pp.307-315
• /
• 2013

Objective: Colistimethate was first became available in 1950s and used until the early 1980s to treat infections caused by gram-negative bacteria and was abandoned due to its nephrotoxicity and neurotoxicity. However, it was recently reintroduced into the clinical practices due to emergence of multidrug-resistance gram-negative bacteria, particularly Pseudomonas aeruginosa and Acinetobacter baumanii. Therefore, it is increasingly used in the intensive care unit settings as a salvage therapy. This study was designed to investigate the incidence rates and risk factors of acute kidney injury associated with colistimethate by using the standardized definition in critically ill patients. Methods: This study retrospectively reviewed the electronic medical records of 71 adult patients above 18 years old receiving intravenous colistimethate at least 48 hours at intensive care unit, university-affiliated hospital from Nov 2012 to Aug 2013 and excluded patients with end-stage renal disease (ESRD) and required renal replacement therapy before initiation of the colistimethate therapy. Acute kidney injury (AKI) was determined by using the standardized RIFLE criteria, classified with risk, injury, failure, loss and ESRD according to serum creatinine (Scr) levels. Results: Among the 71 patients included in the analysis, AKI developed in 40 patients (56.3%) and 6 patients (8.4%) had irreversible kidney injury. AKI occurred within 5 days in 20 patients (50.0%). Maximum Scr level showed a significant increase in the patients with AKI ($1.92{\pm}0.86mg/dL$ vs. $1.12{\pm}0.46mg/dL$ p=0.001), maximum BUN also increased ($64.2{\pm}28.7mg/dL$ vs. $48.4{\pm}24.9mg/dL$ p=0.017) and minimum creatinine clearance (CLcr) was significantly decreased in the patients with AKI than non-AKI ($34.5{\pm}18.6ml/min$ vs. $64.4{\pm}33.7ml/min$ p=0.185). The patients with AKI had significantly longer duration of colistimethate therapy ($21.1{\pm}17.0$ days vs. $13.0{\pm}11.5$ days, p=0.020) and larger cumulative doses of colistimethate ($6465.9{\pm}4717.0mg$ vs. $4438.1{\pm}3426.7mg$, p=0.040). Conclusion: The incidence and severity of AKI associated with colistimethate in critically ill patients was high and serious. Drug monitoring program should be performed to shorten duration of therapy and reduce cumulative dose from initiation of colistimethate therapy for minimizing AKI of colistimethate.