• Journal of Intelligence and Information Systems
• /
• v.25 no.3
• /
• pp.43-62
• /
• 2019

At one time, the anomaly detection sector dominated the method of determining whether there was an abnormality based on the statistics derived from specific data. This methodology was possible because the dimension of the data was simple in the past, so the classical statistical method could work effectively. However, as the characteristics of data have changed complexly in the era of big data, it has become more difficult to accurately analyze and predict the data that occurs throughout the industry in the conventional way. Therefore, SVM and Decision Tree based supervised learning algorithms were used. However, there is peculiarity that supervised learning based model can only accurately predict the test data, when the number of classes is equal to the number of normal classes and most of the data generated in the industry has unbalanced data class. Therefore, the predicted results are not always valid when supervised learning model is applied. In order to overcome these drawbacks, many studies now use the unsupervised learning-based model that is not influenced by class distribution, such as autoencoder or generative adversarial networks. In this paper, we propose a method to detect anomalies using generative adversarial networks. AnoGAN, introduced in the study of Thomas et al (2017), is a classification model that performs abnormal detection of medical images. It was composed of a Convolution Neural Net and was used in the field of detection. On the other hand, sequencing data abnormality detection using generative adversarial network is a lack of research papers compared to image data. Of course, in Li et al (2018), a study by Li et al (LSTM), a type of recurrent neural network, has proposed a model to classify the abnormities of numerical sequence data, but it has not been used for categorical sequence data, as well as feature matching method applied by salans et al.(2016). So it suggests that there are a number of studies to be tried on in the ideal classification of sequence data through a generative adversarial Network. In order to learn the sequence data, the structure of the generative adversarial networks is composed of LSTM, and the 2 stacked-LSTM of the generator is composed of 32-dim hidden unit layers and 64-dim hidden unit layers. The LSTM of the discriminator consists of 64-dim hidden unit layer were used. In the process of deriving abnormal scores from existing paper of Anomaly Detection for Sequence data, entropy values of probability of actual data are used in the process of deriving abnormal scores. but in this paper, as mentioned earlier, abnormal scores have been derived by using feature matching techniques. In addition, the process of optimizing latent variables was designed with LSTM to improve model performance. The modified form of generative adversarial model was more accurate in all experiments than the autoencoder in terms of precision and was approximately 7% higher in accuracy. In terms of Robustness, Generative adversarial networks also performed better than autoencoder. Because generative adversarial networks can learn data distribution from real categorical sequence data, Unaffected by a single normal data. But autoencoder is not. Result of Robustness test showed that he accuracy of the autocoder was 92%, the accuracy of the hostile neural network was 96%, and in terms of sensitivity, the autocoder was 40% and the hostile neural network was 51%. In this paper, experiments have also been conducted to show how much performance changes due to differences in the optimization structure of potential variables. As a result, the level of 1% was improved in terms of sensitivity. These results suggest that it presented a new perspective on optimizing latent variable that were relatively insignificant.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.8 no.4
• /
• pp.145-159
• /
• 2013

The purpose of this study is to analyze the structural relationship as to how learning flow and self-directed learning are linked to learning motives and academic self-efficacy in the learning setting of high school students. To accomplish such purpose, based on theoretical backgrounds and preceding research findings evaluation models were put to verification for a valid research model for this study. The initial hypothetical model was that self-directed learning ability would have a direct influence on learning motive, academic efficacy and learning flow, while having an indirect influence on learning flow with learning motive and self-efficacy acting as a mediating variable. But the hypothetical model showed low significance level between self-directed learning and learning motive, and learning motive and learning flow. Therefore, links were adjusted to create the final model within the scope that the adequacy of the model might not be compromised. To verify the model, 900 high school students in Seoul were surveyed and the collected data were statistically analyzed using AMOS v21.0 and SPSS v21.0 But 815 surveys were excluded because they were not sufficiently answered. From the analysis, it was found that self-directed learning and academic efficacy have a direct influence on learning flow while self-directed learning and academic efficacy have an indirect leaning motive and learning flow. This finding means that, in the relationship of self-directed learning and learning flow, learning motive and learning efficacy are positive factors that help high school students experience learning flow. Thus, in order to enhance the experience of self-directed learning ability of high school students, various educational endeavors are needed to draw the experience of learning flow during the regular course of study. In addition, customized educational methods and environments are required to increase academic efficacy of the students.

• Journal of Korean Society of Environmental Engineers
• /
• v.34 no.8
• /
• pp.566-573
• /
• 2012

This study examined the removal rate of heavy metals from synthetic control water using red mud and lime stone. Overall, the percent of absorption obtained in this study for the red mud treatment was 94.0% ($Pb^{2+}$), 67.1% ($As^{3+}$), 37.5% ($Cu^{2+}$), and 36.6% ($Cr^{6+}$), while that of lime stone was $Pb^{2+}$ (30.8%), $Cu^{2+}$ (16.5%), $Cr^{6+}$ (11.5%), and $As^{3+}$ (8.9%). The kinetic data presented that the slow course of adsorption follows the Pseudo first and second order models, the equilibriuim adsorption of $Cr^{6+}$ and $Pb^{2+}$ obeys Freundlich isotherm model, while the adsorption of $Cu^{2+}$ obeys only Langmuir model. The results also showed that adsorption rate slightly increased with increasing pH from 5 to 9. Interestingly, this trend is similar to results obtained as function of loading amount of red mud. Meanwhile, an unit adsorption rate was slightly decreased. For lime stone, it did not much change in adsorption as function of treatment amount. Consequently, it was concluded that the absorbents can be successfully used the removal of the heavy metals from the aqueous solutions.

• Journal of dental hygiene science
• /
• v.11 no.3
• /
• pp.189-197
• /
• 2011

Objectives : The systematic health education to form the basic healthy lifestyle should be realized from elementary school, so oral health education at elementary school can determine the whole lifetime oral health. The elementary school health teacher's recognition and behavior who in charge of health promotion of students is important. Therefore, the study was conducted to enhance oral health education. Methods : Total 114 people among of 131 from health teacher Gwangju elementary school. Survey system is configured by referring to PRECEDE model. Results : In behavioral diagnosis the proportion of oral Health Education is less than 10%(58.8%), mostly educated in activity time (86.0%), the health teachers educate when it is needed(53.5%). In predisposing diagnosis in the eight areas of health education, the oral health education is ranked as fourth, fifth. Even in the next year project, the oral health education ratio was 21.9 percent. In enabling diagnosing every year the Oral health education training experienced rate is 13.2%, satisfaction rate is 33.3%. In reinforcing diagnosing disability element in the regular education course are the lack of oral health-related information (46.7%), lack of materials needed for education (30.6%), lack of training opportunities (21.4%). Conclusion : In further research, oral health education textbooks, materials and methods should be developed. At the foundation of there developments, the elementary oral health education program should be more fully developed and conducted and also the evaluation of its effectiveness will need.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.32 no.3B
• /
• pp.175-183
• /
• 2012

In this study, Galerkin scheme and SU/PG scheme of Petrov-Galerkin family were applied to the shallow water equations and a finite element model for shallow water flow was developed. Numerical simulations were conducted in several flumes with convection-dominated flow condition. Flow simulation of channel with slender structure in the water course revealed that Galerkin and SU/PG schemes showed similar results under very low Fr number and Re number condition. However, when the Fr number increased up to 1.58, Galerkin scheme did not converge while SU/PG scheme produced stable solutions after 5 iterations by Newton-Raphson method. For the transcritical flow simulation in diverging channel, the present model predicted the hydraulic jump accurately in terms of the jump location, the depth slope, and the flow depth after jump, and the numerical results showed good agreements with the hydraulic experiments carried out by Khalifa(1980). In the oblique hydraulic jump simulation, in which convection-dominated supercritical flow (Fr=2.74) evolves, Galerkin scheme blew up just after the first iteration of the initial time step. However, SU/PG scheme captured the boundary of oblique hydraulic jump accurately without numerical oscillation. The maximum errors quantified with exact solutions were less than 0.2% in water depth and velocity calculations, and thereby SU/PG scheme predicted the oblique hydraulic jump phenomena more accurately compared with the previous studies (Levin et al., 2006; Ricchiuto et al., 2007).

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.1
• /
• pp.149-154
• /
• 2014

Quality is the value that can be measured by observing the characteristics of the service quantity or quality. QoS is predictable service traffic to a minimum requirements what passed in network. In the course of Smart Medical Information System Development there exist some functional requirements to satisfy quality objectives. The functional smart domains of healthcare information systems consists of Patient Module, a smart sensing and communication domain, RFID Tag Readers and the behavior domain, Homecare Station Domain, Clinical Station. This study is performed on evaluation methodology of u-health service satisfaction quality of each domain. In this paper QoS metrics and the quality of medical information requirements, functional requirements are separated by. Quality parameters consists of six items and the functional requirements and quality requirements 20 details the five items and consist of 20 detailed items. On this study the quality evaluation methodology of Korean smart health information quality assessment matrix 2 - factor evaluation method is proposed. The overall framework of this paper is organizing the specific criteria of quality of medical information system and modeling quality evaluation process under all smart environment.

• Journal of Ocean Engineering and Technology
• /
• v.33 no.6
• /
• pp.590-596
• /
• 2019

Numerous deaths and substantial property damage have occurred recently due to frequent disasters of the highest intensity according to the abnormal climate, which is caused by various problems, such as global warming, all over the world. Such large-scale disasters have become an international issue and have made people aware of the disasters so they can implement disaster-prevention measures. Extensive information on disaster prevention actively has been announced publicly to support the natural disaster reduction measures throughout the world. In Japan, diverse developmental studies on disaster prevention systems, which support hazard map development and flood control activity, have been conducted vigorously to estimate external forces according to design frequencies as well as expected maximum frequencies from a variety of areas, such as rivers, coasts, and ports based on broad disaster prevention data obtained from several huge disasters. However, the current reduction measures alone are not sufficiently effective due to the change of the paradigms of the current disasters. Therefore, in order to obtain the synergy effect of reduction measures, a study of the establishment of an integrated system is required to improve the various disaster prevention technologies and the current disaster prevention system. In order to develop a similar typhoon search system and establish a disaster prevention infrastructure, in this study, techniques will be developed that can be used to forecast typhoons before they strike by using artificial intelligence (AI) technology and offer primary disaster prevention information according to the direction of the typhoon. The main function of this model is to predict the most similar typhoon among the existing typhoons by utilizing the major typhoon information, such as course, central pressure, and speed, before the typhoon directly impacts South Korea. This model is equipped with a combination of AI and DNN forecasts of typhoons that change from moment to moment in order to efficiently forecast a current typhoon based on similar typhoons in the past. Thus, the result of a similar typhoon search showed that the quality of prediction was higher with the grid size of one degree rather than two degrees in latitude and longitude.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.10 no.1
• /
• pp.53-60
• /
• 1981

Dehydration phenomena has been studied for the shiitake mushroom Lentinus edodes sanryun No.1, through which examine the effect of temperature and air velocity and derivation of its kinetics. Temperature effect for the dehydration rate constant were examined under the constant air velocity (1.5m/sec) with the variation of temperature from $40^{\circ}C$ to $55^{\circ}C$. Water content were reduced exponentially with the course of time and calculated dehydration rate constant values varies with temperature with an Arrhenius-type relationship, which had been expected in the chemical reaction kinetics. Influence of air velocity for the dehydration rate constant under the constant temperature $(45^{\circ}C)$ showed interesting results. For the range 1.0m/sec to 2.0m/sec, dehydration rate constant values are increased with the air velocity, but for the 2.0 to 3.1m/sec, dehydration rate constant values are decreased which were caused by case hardening. One of the selected conditions in the optimal dehydration range, temperature $50^{\circ}C$, air velocity 2m/see, and its measured humidity 38-41%, mathematical model of dehydration curve and dehydration rate equations were developed and the resulting kinetic models were X=6.94 $e^{-0.345t}$ and dx/dt = -2.39 $e^{-0.345t}$

• Horticultural Science & Technology
• /
• v.34 no.2
• /
• pp.236-241
• /
• 2016

The common ice plant (Mesembryanthemum crystallinum L.) has some medicinal uses and recommended plant in closed-type plant factory. The objective of this study was to estimate the cardinal temperatures for seed germination of the common ice plant using bilinear, parabolic, and beta distribution models. Seeds of the common ice plant were germinated in the dark in a growth chamber at four constant temperatures: 16, 20, 24, and $28^{\circ}C$. For this, four replicates of 100 seeds were placed on two layers of filter paper in a 9-cm petri dish and radicle emergence of 0.1 mm was scored as germination. The times to 50% germination were 4.3, 2.5, 2.0, and 1.8 days at 16, 20, 24, and $28^{\circ}C$, respectively, indicating that the germination of this warm-weather crop increased with temperature. Next, the time course of germination was modeled using a logistic function. For the selection of an accurate model, seeds were germinated in the dark at constant temperatures of 6, 12, 32, and $36^{\circ}C$. Germination started earlier and increased rapidly at temperatures above $20^{\circ}C$. The minimum, optimal, and maximum temperatures were estimated by regression of the inverse of time to 50% germination rate, as a function of the temperature gradient. The different functions estimated differing minimum, optimal and maximum temperatures, with 5.7, 27.7, and $36.5^{\circ}C$, respectively for the bilinear function, 13.4, 25.0, and $36.6^{\circ}C$, respectively, for the parabolic function and 7.8, 25.9, and $36.0^{\circ}C$, respectively, for the beta distribution function. The models estimated that the inverse of time to 50% germination rate was 0 at 6 and $36^{\circ}C$. The observed final germination rates at 12 and $32^{\circ}C$ were 62 and 97%, respectively. Our data show that a beta distribution function provides a useful model for estimating the cardinal temperatures for germination of seed from the common ice plant.

• Journal of Korean Neurosurgical Society
• /
• v.54 no.6
• /
• pp.467-476
• /
• 2013

Objective : This study aimed to evaluate the hypotheses that administration routes [intra-arterial (IA) vs. intravenous (IV)] affect the early stage migration of transplanted human bone marrow-derived mesenchymal stem cells (hBM-MSCs) in acute brain infarction. Methods : Male Sprague-Dawley rats (n=40) were subjected to photothrombotic infarction. Three days after photothrombotic infarction, rats were randomly allocated to one of four experimental groups [IA group : n=12, IV group : n=12, superparamagnetic iron oxide (SPIO) group : n=8, control group : n=8]. All groups were subdivided into 1, 6, 24, and 48 hours groups according to time point of sacrifice. Magnetic resonance imaging (MRI) consisting of T2 weighted image (T2WI), $T2^*$ weighted image ($T2^*WI$), susceptibility weighted image (SWI), and diffusion weighted image of rat brain were obtained prior to and at 1, 6, 24, and 48 hours post-implantation. After final MRI, rats were sacrificed and grafted cells were analyzed in brain and lung specimen using Prussian blue and immunohistochemical staining. Results : Grafted cells appeared as dark signal intensity regions at the peri-lesional zone. In IA group, dark signals in peri-lesional zone were more prominent compared with IV group. SWI showed largest dark signal followed by $T2^*WI$ and T2WI in both IA and IV groups. On Prussian blue staining, IA administration showed substantially increased migration and a large number of transplanted hBM-MSCs in the target brain than IV administration. The Prussian blue-positive cells were not detected in SPIO and control groups. Conclusion : In a rat photothrombotic model of ischemic stroke, selective IA administration of human mesenchymal stem cells is more effective than IV administration. MRI and histological analyses revealed the time course of cell migration, and the numbers and distribution of hBM-MSCs delivered into the brain.