• Journal of The Korean Astronomical Society
• /
• v.52 no.6
• /
• pp.217-225
• /
• 2019

We apply a modified Convolutional Neural Network (CNN) model in conjunction with transfer learning to predict whether an active region (AR) would produce a ≥C-class or ≥M-class flare within the next 24 hours. We collect line-of-sight magnetogram samples of ARs provided by the SHARP from May 2010 to September 2018, which is a new data product from the HMI onboard the SDO. Based on these AR samples, we adopt the approach of shuffle-and-split cross-validation (CV) to build a database that includes 10 separate data sets. Each of the 10 data sets is segregated by NOAA AR number into a training and a testing data set. After training, validating, and testing our model, we compare the results with previous studies using predictive performance metrics, with a focus on the true skill statistic (TSS). The main results from this study are summarized as follows. First, to the best of our knowledge, this is the first time that the CNN model with transfer learning is used in solar physics to make binary class predictions for both ≥C-class and ≥M-class flares, without manually engineered features extracted from the observational data. Second, our model achieves relatively high scores of TSS = 0.640±0.075 and TSS = 0.526±0.052 for ≥M-class prediction and ≥C-class prediction, respectively, which is comparable to that of previous models. Third, our model also obtains quite good scores in five other metrics for both ≥C-class and ≥M-class flare prediction. Our results demonstrate that our modified CNN model with transfer learning is an effective method for flare forecasting with reasonable prediction performance.

• Cross-Cultural Studies
• /
• v.25
• /
• pp.549-596
• /
• 2011

This paper aims to introduce some useful multimedia freewares, and also support Chinese instructor with discussing how to operate them for instructional design of multimedia language learning class. For this aims, this paper consists of three parts: First, instructional design. This part is focused to what kind of instructional model to be based on, for example, Dick & Carey model, ADDIE model, ASSURE model etc. This part introduces these models, and modifies ADDIE and ASSURE model to D.D.A.I.E.S and S.S.A.U.R.E.S as 'A(nalysis)' in these model may apply to the next 'D(evelopment)' on ADDIE, 'S(elect Methods, Media and Materials)' on ASSURE in the practical Chinese class. Second, Programme: What to use. This part is focused to what kind of free software we can use. In the web site online, there are huge free softwares so we usually hesitate to select and also don't know how to operate even though selected one of them. This part, accordingly, introduces ten of useful freewares and compares each other in terms of usefulness for Chinese instructors. Third, Programme: How to use. It is of no use just to know what to use but not to know how to operate, so this part describes how to use freewares like a kind of manual in detail as far as possible. In conclusion, we hope more Chinese instructors to learn and use more useful freewares for designing the better multimedia Chinese class by this paper.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.38 no.1
• /
• pp.139-152
• /
• 2013

The purpose of this study is to claim the validity of tuning the architecture of neural network models for multi-class classification. A neural network model for multi-class classification is basically constructed by building a series of neural network models for binary classification. Building a neural network model, we are required to set the values of parameters such as number of hidden nodes and weight decay parameter in advance, which draws special attention as the performance of the model can be quite different by the values of the parameters. For better performance of the model, it is absolutely necessary to have a prior process of tuning the parameters every time the neural network model is built. Nonetheless, previous studies have not mentioned the necessity of the tuning process or proved its validity. In this study, we claim that we should tune the parameters every time we build the neural network model for multi-class classification. Through empirical analysis using wine data, we show that the performance of the model with the tuned parameters is superior to those of untuned models.

• The korean journal of orthodontics
• /
• v.45 no.3
• /
• pp.113-120
• /
• 2015

Objective: This study compared occlusal contact areas of ideally planned set-up and accomplished final models against the initial in class I and II molar relationships at finishing. Methods: Evaluations were performed for 41 post-orthodontic treatment cases, of which 22 were clinically diagnosed as class I and the remainder were diagnosed as full cusp class II. Class I cases had four first premolars extracted, while class II cases had maxillary first premolars extracted. Occlusal contact areas were measured using a three-dimensional scanner and RapidForm 2004. Independent t-tests were used to validate comparison values between class I and II finishings. Repeated measures analysis of variance was used to compare initial, set up, and final models. Results: Molars from cases in the class I finishing for the set-up model showed significantly greater contact areas than those from class II finishing (p < 0.05). The final model class I finishing showed significantly larger contact areas for the second molars (p < 0.05). The first molars of the class I finishing for the final model showed a tendency to have larger contact areas than those of class II finishing, although the difference was not statistically significant (p = 0.078). Conclusions: In set-up models, posterior occlusal contact was better in class I than in class II finishing. In final models, class I finishing tended to have larger occlusal contact areas than class II finishing.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.1
• /
• pp.239-246
• /
• 2019

In this paper, we propose an effective software education donation model for the social care class. The types of software education for elementary, middle, and high school for the social care class are in the order of after school classes, club activities, creative experiences, and regular classes. In elementary school students, it is effective to precede visual programming education based on block coding and to conduct curriculum convergence with SW and HW at the beginning, and high school students are carrying out text programming education like Python. Software education for social care class The contribution activity model can be classified into five types such as geographically difficult area, multicultural family areas, orphanage, reformatory, and basic livelihood security recipient. In addition, the survey results show that the students' interest in software education and their satisfaction are all very high at 96%. Effective software education for the social care class In the donation model, the lecturers consist of responsible professors, lecturers, and assistant instructors. Software training for the social care class is effective on a year-by-year basis, so that students can feel authenticity and trust. Software education contents focus on visual programming and physical computing education in elementary or middle school, and text programming and physical computing education in high school. It is necessary to construct a software education donor matching system that helps efficient management of software education donations by efficiently matching schools (consumers: elementary, middle, high school) and software education donors(suppliers).

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.12
• /
• pp.3668-3674
• /
• 2009

Use case diagram is used for defining inter-action between users and systems in object-oriented modeling. It is very important to defining users' requirements for efficient software development. In this paper, we propose a object-oriented modeling process using hierarchical use case analysis for designing class model. First, We define many use case diagrams by several hierarchical modeling level. And next, we can also design class model using the use case diagrams. Our proposed modeling process provides interaction between use case model and class model. That can make us to check the modeling process during the software development. Using the proposed object-oriented modeling we can develop software based on users' requirements. It is very useful for class modeling.

• Tunnel and Underground Space
• /
• v.21 no.3
• /
• pp.216-224
• /
• 2011

In this study, Class II behavior of rock failure process under uniaxial and biaxial compression has been numerically simulated using bonded particle model. Class II behavior of rock was simulated by radial strain controlled uniaxial and biaxial compression tests using a suggested method of ISRM. Micro-parameters used in the simulation were determined based on the laboratory uniaxial compression tests carried out at ${\"{A}}sp{\"{o}}$ Hard Rock Laboratory, Sweden. Class II behavior of ${\"{A}}sp{\"{o}}$ rock was effectively simulated using newly proposed numerical technique in this study, and the results of numerical simulations show good similarity with the complete stress-strain curves for Class II behavior obtained from the laboratory tests.

• Journal of Fisheries and Marine Sciences Education
• /
• v.28 no.5
• /
• pp.1478-1491
• /
• 2016

This study has designed to find out the effects of a project-based flipped learning model at universities on the self-directed learning ability, the self-leadership, and the learning competency. For the study, two procedures were performed. First, a flipped learning model for a project-based learning was developed on the basis of the literature reviews. The flipped learning model has three different steps: the pre-class, the in-class, and the post-class. In the pre-class, instructors provide mini-core courses using various technologies for learners outside the class. The in-class is the step to check whether learners prepare their learning or not. Also, in this step, the in-death learning and the teaching-learning process by interaction between instructors and learners would be performed. In the post-class, learners would be able to sustain the extended learning to develop the learning tasks and activities after flipped learning class. Through this step, the learners could be experienced integrated thinking and application, documentation and management, as well as sharing and spread of their learning. Second, the effectiveness of the developed flipped learning model on the self-direction, the self-leadership, and the learning competency was examined. The quantitative research method and the qualitative research method were used for this study. The results indicated that the flipped learning model showed improvement on the self-direction, the self-leadership, and the learning competency.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.6
• /
• pp.15-21
• /
• 2018

The purpose of this study is to design a Content-based Korean Class model using Flipped learning for foreign students. The class model that presents on this paper will lead the language learning through content learning, also it will be enable the student more active and to have an initiative in the class. Prior to designing a Content-based Korean Class model using Flipped learning, the concepts and educational significance and characteristics of flip learning were reviewed through previous studies. Then, It emphasizes the necessity of teaching method adapting Flipped learning to Content-based teaching method in Korean language education. It also suggests standards and principles of composition in Contents-based teaching method using Flipped learning. After designing the instructional model based on the suggested standards and principles, it presents a course of instruction about how learning methods, contents and activities should be done step by step. The Content-based Korean class model using the Flipped learning will be an alternative approach to overcome the limitations of teacher-centered teaching methods and lecture-teaching methods which are the dominant of present classroom environment.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 1999.11a
• /
• pp.177-183
• /
• 1999

In this paper, we analyze transmission characteristics of ultrashort laser pulses using the property of Raised-cosine pulse which are systematically obtained following Class-2 PRS model. The high-order pulses are easily derived front a modified PRS system model as Class-1 PRS model. This may be based on the fact that the spectra and bandwidths of the high-order pulses are beautifully related to their orders. And we make clear they are very useful to cover wider area and more accurate transmission characteristics of ultrashort pulses than Gaussian or Sech pulse approximations used conventionally. First modifying the generalized PRS system model, we propose a new model for deriving any type of high-order pulse. And we offer a novel analysis method of ultrashort pulse transmission which has any shape and FWHM, using the proposed model. In addition, by fixing the pulse range $\tau$=1(ps) and varying the order of the pulse from n=1 to n=100, we obtain spectra of ultrashort pulses with 1(ps)-100(fs) FWHM's, and width of FWHM in the Class-2 PRS model 50~100(fs) smaller than Class-1 PRS model. As a one-step further, we derive PSD's of their pulse trains when they are applied to Unipolar signaling scheme. These PSD's are derided in the range of possible pulse intervals. All of these results are not only coincided with some conventional experimental works but also will to applied to any pioneering ultrashort pulse in the future.