• The Journal of the Acoustical Society of Korea
• /
• v.34 no.3
• /
• pp.240-246
• /
• 2015

In this paper, speech enhancement using nonnegative matrix factorization with temporal continuity has been addressed. Speech and noise signals are modeled as Possion distributions, and basis vectors and gain vectors of NMF are modeled as Gamma distributions. Temporal continuity of the gain vector is known to be critical to the quality of enhanced speech signals. In this paper, temporal continiuty is implemented by adopting Gamma-Markov chain priors for noise gain vectors during the separation phase. Simulation results show that the Gamma-Markov chain models temporal continuity of noise signals and track changes in noise effectively.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.3 s.303
• /
• pp.1-6
• /
• 2005

In the Markov network which models disparity map with the Markov Random Fields(MRF), the belief propagation algorithm is operated by message passing between nodes corresponding to each pixel. The initial message value is converged by iterations of the algorithm and the algorithm requires many iterations to get converged messages. In this paper, we simplify the algorithm by regarding the objects in the disparity map as combinations of lines with same message valued nodes to reduce iterations of the algorithm.

• Journal of the Korea Society of Computer and Information
• /
• v.17 no.5
• /
• pp.33-40
• /
• 2012

In this paper we propose an anomaly detection scheme to detect new attack paths or new attack methods without false positives by monitoring HTTP Outbound Traffic after efficient training. Our proposed scheme detects web-based attacks by comparing tags or javascripts of HTTP Outbound Traffic with normal behavioral models which apply HMM(Hidden Markov Model). Through the verification analysis under the real-attacked environment, we show that our scheme has superior detection capability of 0.0001% false positive and 96% detection rate.

• Journal of the Korea Computer Graphics Society
• /
• v.15 no.2
• /
• pp.9-17
• /
• 2009

Animating multiple characters to compete with each other is an important problem in computer games and animation films. However, it remains difficult to simulate strategic competition among characters because of its inherent complex decision process that should be able to cope with often unpredictable behavior of opponents. We adopt a reinforcement learning method in Markov games to action models built from captured motion data. This enables two characters to perform globally optimal counter-strategies with respect to each other. We also extend this method to simulate competition between two teams, each of which can consist of an arbitrary number of characters. We demonstrate the usefulness of our approach through various competitive scenarios, including playing-tag, keeping-distance, and shooting.

• IE interfaces
• /
• v.16 no.spc
• /
• pp.21-27
• /
• 2003

Objective of this paper is to develop the user mobility model(UMM) which is used for the performance analysis of location update and paging algorithm and at the same time, consider the user mobility pattern(UMP) in zone-based cellular networks. User mobility pattern shows correlation in space and time. UMM should consider these correlations of UMP. K-dimensional Markov chain is presented as a UMM considering them where the states of Markov chain are defined as the current location area(LA) and the consecutive LAs visited in the path. Also, a new two dimensional Markov chain composed of current LA and time interval is presented. Simulation results show that the appropriate size of K in the former UMM is two and the latter UMM reflects the characteristic of UMP well and so is a good model for the analytic method to solve the performance of location update and paging algorithm.

• Proceedings of the KAIS Fall Conference
• /
• 2003.11a
• /
• pp.141-148
• /
• 2003

Customer retention is a common concern for many industries and a critical issue for the survival in today's greatly compressed marketplace. Current customer retention models only focus on detection of potential defectors based on the likelihood of defection by using demographic and customer profile information. In this paper, we propose a dynamic procedure for defection detection and prevention using past and current customer behavior by utilizing SOM and Markov chain. The basic idea originates from the observation that a customer has a tendency to change his behavior (i.e. trim-out his usage volumes) before his eventual withdrawal. This gradual pulling out process offers the company the opportunity to detect the defection signals. With this approach, we have two significant benefits compared with existing defection detection studies. First, our procedure can predict when the potential defectors could withdraw and this feature helps to give marketing managers ample lead-time for preparing defection prevention plans. The second benefit is that our approach can provide a procedure for not only defection detection but also defection prevention, which could suggest the desirable behavior state for the next period so as to lower the likelihood of defection. We applied our dynamic procedure for defection detection and prevention to the online gaming industry. Our suggested procedure could predict potential defectors without deterioration of prediction accuracy compared to that of the MLP neural network and DT.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.25 no.4
• /
• pp.319-325
• /
• 2007

In case of more than 50km baseline length, the correlation between receivers is reduced. Therefore, there are still some rooms for improvement of its positional accuracy. In this paper, the stochastic modeling of the ionospheric delay is applied and its effects are analyzed. The data processing has been performed by constructing a Kalman filter with states of positions, ambiguities, and the ionospheric delays in the double differenced mode. Considering the medium or long baseline length, both double differenced GPS phase and code observations are used as observables and LAMBDA has been applied to fix the ambiguities. The ionospheric delay is stochastically modeled by well-known 1st order Gauss-Markov process. And the correlation time and variation of 1st order Gauss-Markov process are calculated. This paper gives analyzed results of developed algorithm compared with commercial software and Bernese.

• Journal of Korean Society for Atmospheric Environment
• /
• v.19 no.4
• /
• pp.397-414
• /
• 2003

The performance of one-particle stochastic Lagrangian models for passive tracer dispersion are evaluated against measurements in horizontally-homogeneous neutrally-stratified atmospheric surface layer. State-of-the-technology models as well as classical Langevin models, all in class of well mixed models are numerically implemented for inter-model comparison study. Model results (far-downstream asymptotic behavior and vertical profiles of the time averaged concentrations, concentration fluxes, and concentration fluctuations) are compared with the reported measurements. The results are: 1) the far-downstream asymptotic trends of all models except Reynolds model agree well with Garger and Zhukov's measurements. 2) profiles of the average concentrations and vertical concentration fluxes by all models except Reynolds model show good agreement with Raupach and Legg's experimental data. Reynolds model produces horizontal concentration flux profiles most close to measurements, yet all other models fail severely. 3) With temporally correlated emissions, one-particle models seems to simulate fairly the concentration fluctuations induced by plume meandering, when the statistical random noises are removed from the calculated concentration fluctuations. Analytical expression for the statistical random noise of one-particle model is presented. This study finds no indication that recent models of most delicate theoretical background are superior to the simple Langevin model in accuracy and numerical performance at well.

• Journal of Industrial Technology
• /
• v.28 no.A
• /
• pp.185-189
• /
• 2008

Practical pattern recognition systems should overcome very large class problem. Sometimes it is almost impossible to build every model for every class due to memory and time constraints. For this case, grouping similar models will be helpful. In this paper, we propose GMC(Group Model Clustering) to build a large class Chinese character recognition system. We built hidden Markov models for 10% of total classes, then classify the rest of classes into already trained group classes. Finally group models are trained using group model clustered data. Recognition is performed using only group models, in order to achieve reduced model size and improved recognition speed.

• Journal of Korean Society for Quality Management
• /
• v.27 no.3
• /
• pp.125-141
• /
• 1999

Bayesian inference and model selection method for software reliability growth models are studied. Software reliability growth models are used in testing stages of software development to model the error content and time intervals between software failures. In this paper, we could avoid the multiple integration by the use of Gibbs sampling, which is a kind of Markov Chain Monte Carlo method to compute the posterior distribution. Bayesian inference and model selection method for Jelinski-Moranda and Goel-Okumoto and Schick-Wolverton models in software reliability with Poisson prior information are studied. For model selection, we explored the relative error.