• Journal of the Korea Society for Simulation
• /
• v.3 no.1
• /
• pp.65-74
• /
• 1994

In the steady-state simulation, it is important to identify initialization bias for the correct estimates of the simulation model under study. In this paper, the methods from chaos theory are applied to the determination of truncation points in the simulation data for controlling the initial bias. Two methods are proposed and evaluated based on their effectiveness for estimation the average waiting time in M/M/1($\infty$) queueing model.

• Journal of the Korean earth science society
• /
• v.35 no.5
• /
• pp.342-353
• /
• 2014

Data for model analysis derived from the finite volume (fv) GCM (Goddard Earth Observing System Ver. 4, GEOS-4) and the Land Data Assimilation System (LDAS) have been utilized in a mesoscale model. These data are tested to provide initial conditions and lateral boundary forcings to the Purdue Mesoscale Model (PMM) for a case study of the Midwestern flood that took place from 21-23 May 1998. The simulated results with fvGCM and LDAS soil moisture and temperature data are compared with that of ECMWF reanalysis. The initial conditions of the land surface provided by fvGCM/LDAS show significant differences in both soil moisture and ground temperature when compared to ECMWF control run, which results in a much different atmospheric state in the Planetary Boundary Layer (PBL). The simulation result shows that significant changes to the forecasted weather system occur due to the surface initial conditions, especially for the precipitation and temperature over the land. In comparing precipitation, moisture budgets, and surface energy, not only do the intensity and the location of precipitation over the Midwestern U.S. coincide better when running fvGCM/LDAS, but also the temperature forecast agrees better when compared to ECMWF reanalysis data. However, the precipitation over the Rocky Mountains is too large due to the cumulus parameterization scheme used in the PMM. The RMS errors and biases of fvGCM/LDAS are smaller than the control run and show statistical significance supporting the conclusion that the use of LDAS improves the precipitation and temperature forecast in the case of the Midwestern flood. The same method can be applied to Korea and simulations will be carried out as more LDAS data becomes available.

• Atmosphere
• /
• v.23 no.1
• /
• pp.73-83
• /
• 2013

In recent several years, East Asia, Europe and North America have suffered successive cold winters and a number of historical records on the extreme weathers are replaced with new record-breaking cold events. As a possible explanation, several studies suggested that cryospheric conditions of Northern Hemisphere (NH), i.e. Arctic sea-ice and snow cover over northern part of major continents, are changing significantly and now play an active role for modulating midlatitude atmospheric circulation patterns that could bring cold winters for some regions in midlatitude. In this study, a dynamical seasonal prediction system for NH winter is newly developed using the snow depth initialization technique and statistically predicted sea-ice boundary condition. Since the snow depth shows largest variability in October, entire period of October has been utilized as a training period for the land surface initialization and model land surface during the period is continuously forced by the observed daily atmospheric conditions and snow depths. A simple persistent anomaly decaying toward an averaged sea-ice condition has been used for the statistical prediction of sea-ice boundary conditions. The constructed dynamical prediction system has been tested for winter 2012/13 starting at November 1 using 16 different initial conditions and the results are discussed. Implications and a future direction for further development are also described.

• International Journal of Computer Science & Network Security
• /
• v.23 no.8
• /
• pp.177-189
• /
• 2023

Malware detection is an increasingly important operational focus in cyber security, particularly given the fast pace of such threats (e.g., new malware variants introduced every day). There has been great interest in exploring the use of machine learning techniques in automating and enhancing the effectiveness of malware detection and analysis. In this paper, we present a deep recurrent neural network solution as a stacked Long Short-Term Memory (LSTM) with a pre-training as a regularization method to avoid random network initialization. In our proposal, we use global and short dependencies of the inputs. With pre-training, we avoid random initialization and are able to improve the accuracy and robustness of malware threat hunting. The proposed method speeds up the convergence (in comparison to stacked LSTM) by reducing the length of malware OpCode or bytecode sequences. Hence, the complexity of our final method is reduced. This leads to better accuracy, higher Mattews Correlation Coefficients (MCC), and Area Under the Curve (AUC) in comparison to a standard LSTM with similar detection time. Our proposed method can be applied in real-time malware threat hunting, particularly for safety critical systems such as eHealth or Internet of Military of Things where poor convergence of the model could lead to catastrophic consequences. We evaluate the effectiveness of our proposed method on Windows, Ransomware, Internet of Things (IoT), and Android malware datasets using both static and dynamic analysis. For the IoT malware detection, we also present a comparative summary of the performance on an IoT-specific dataset of our proposed method and the standard stacked LSTM method. More specifically, of our proposed method achieves an accuracy of 99.1% in detecting IoT malware samples, with AUC of 0.985, and MCC of 0.95; thus, outperforming standard LSTM based methods in these key metrics.

• Journal of the Korea Society of Computer and Information
• /
• v.3 no.2
• /
• pp.57-67
• /
• 1998

In order to extract the contour of interesting object in the image, Kass suggested the Active Contour Model called "Snakes" The speed of this model is slow and this model is sensitive of initialization. In order to improve these problems, Gunn extracted the accurate contour by using two initializations, and operated to less sensitive of initialization. This method could extract more accurate contour than the existing method, but it had no effect in the speed and it was sensitive of noise. This paper applied to the Energy Minimization Algorithm about only the pixel within the window applying the window of $8{\times}8$ size at each contour point consisting Snakes in order to solve these problems. In order to less sensitive of noise which exists within image, it suggests a method that moves the window to vertical direction for the gradient of each contour point.our point.

• Journal of Microbiology and Biotechnology
• /
• v.14 no.6
• /
• pp.1163-1169
• /
• 2004

An unstructured mathematical model is presented for lactic acid fermentation based on the energy balance. The proposed model reflects the energy metabolic state and then predicts the cell growth, lactic acid production, and glucose consumption rates by relating the above rates with the energy metabolic rate. Fermentation experiments were conducted under various initial lactic acid concentrations of 0, 30, 50, 70, and 90 g/l. Also, a genetic algorithm was used for further optimization of the model parameters and included the operations of coding, initialization, hybridization, mutation, decoding, fitness calculation, selection, and reproduction exerted on individuals (or chromosomes) in a population. The simulation results showed a good fit between the model prediction and the experimental data. The genetic algorithm proved to be useful for model parameter optimization, suggesting wider applications in the field of biological engineering.

• Journal of Biomedical Engineering Research
• /
• v.21 no.5
• /
• pp.457-467
• /
• 2000

In this paper, we propose the method to extract the anatomical objects in medical images using active contour models and genetic algorithms. The performance of active contour models is mostly decided by the optimization of active contour model's energy. So, we propose to use genetic algorithms to optimize the energy of active contour models. We experimented our proposed method on the femoral head medical images and proved that our method provides very acceptable results from any initialization of active contour models.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.3
• /
• pp.141-149
• /
• 2005

Tn general. Gaussian mixture model (GMM) is used to estimate the speaker model from the speech for speaker identification. The parameter estimates of the GMM are obtained by using the Expectation-Maximization (EM) algorithm for the maximum likelihood (ML) estimation. However the EM algorithm has such drawbacks that it depends heavily on the initialization and it needs the number of mixtures to be known. In this paper, to solve the above problems of the EM algorithm. we propose an EM algorithm with the initialization based on incremental ${\cal}k-means$ for GMM. The proposed method dynamically increases the number of mixtures one by one until finding the optimum number of mixtures. Whenever adding one mixture, we calculate the mutual relationship between it and one of other mixtures respectively. Finally. based on these mutual relationships. we can estimate the optimal number of mixtures which are statistically independent. The effectiveness of the proposed method is shown by the experiment for artificial data. Also. we performed the speaker identification by applying the proposed method comparing with other approaches.

• International Journal of Ocean Engineering and Technology Speciallssue:Selected Papers
• /
• v.6 no.1
• /
• pp.22-29
• /
• 2003

This research aims at investigation of accuracy potential of RTK(Real-Time Kinematic) GPS in combination with Echo Sounder(E/S) for the coastal mapping. Apart from this purpose, the accuracy of ambiguity resolution with the OTF(On The Fly) method was tested with respect to the initialization time. The result shows that the accuracy is better than 1cm with 5-minute initialization in the distance of 10km baseline. The seaside topography was measured by the RTK GPS only, on the other hand the seafloor topography was surveyed in combination of RTK GPS and E/S. Comparing to the volume of seaside measured by RTK GPS and digital topographical map, the difference of only 2% was achieved. This indicates that the coastal mapping. As a result, it has been revealed that every possible noise in surveying could be corrected and the accuracy could be improved. The accuracy of GPS data acquired in real time was as good as that acquired by post processing. It is expected that it will be useful for the analysis of coastal geographic characteristics because DTM(Digital Terrain Model) can be also constructed for the harbor reclamation, the dredging, and the variation of soil movement in a river.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2715-2717
• /
• 2002

Image segmentation is an essential technique of image analysis. In spite of the issues in contour initialization and boundary concavities, active contour models(snakes) are popular and successful methods for the segmentation. In this paper, we present a new active contour model, GGF snake, for segmentation of endoscopic image. The GGF snake is less sensitive to contour initialization and ensures high accuracy, large capture range, and fast CPU time for computing external force. It was observed that the GGF snake produced more reasonable results in various image types, such as simple synthetic images, commercial digital camera images, and endoscopic images than previous snakes did.