• 정보관리학회지
• /
• 제19권4호
• /
• pp.35-51
• /
• 2002

본 논문에서는 다중 모델 기계학습 기법을 이용하여 자동 문서 분류의 성능과 신뢰도를 향상시킬 수 있는 연구와 실험 결과를 기술하였다. 기존의 다중 모델 기계 학습법들이 훈련 데이터 또는 학습 알고리즘의 편향에 의한 오류를 극복하고자 한 것인데 비해 본 논문에서 제안한 메타 학습을 이용한 하이브리드 다중 모델 방식은 이 두 가지의 오류 원인을 동시에 해소하고자 하였다. 다양한 문서 집합에 대한 실험 결과. 본 논문에서 제안한 하이브리드 다중 모델 학습법이 전반적으로 기존의 일반 다중모델 학습법들에 비해 높은 성능을 보였으며, 다중 모델의 결합 방식으로서 메타 학습이 투표 방식에 비해 효율적인 것으로 나타났다.

• Journal of information and communication convergence engineering
• /
• 제15권1호
• /
• pp.7-13
• /
• 2017

We consider a distributed estimation where many nodes remotely placed at known locations collect the measurements of the parameter of interest, quantize these measurements, and transmit the quantized data to a fusion node; this fusion node performs the parameter estimation. Noting that quantizers at nodes should operate in a non-regular framework where multiple codewords or quantization partitions can be mapped from a single measurement to improve the system performance, we propose a low-weight estimation algorithm that finds the most feasible combination of codewords. This combination is found by computing the weighted sum of the possible combinations whose weights are obtained by counting their occurrence in a learning process. Otherwise, tremendous complexity will be inevitable due to multiple codewords or partitions interpreted from non-regular quantized data. We conduct extensive experiments to demonstrate that the proposed algorithm provides a statistically significant performance gain with low complexity as compared to typical estimation techniques.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.729-734
• /
• 2005

One of the important research issues in mobile robot is how to detect the 3D environment fast and accurately, and recognize it. Sensing methods of utilizing laser structured light and/or stereo vision are representatively used among a number of methodologies developed to date. However, the methods are still in need of achieving high accuracy and reliability to be used for real world environments. In this paper to implement a new robotic environmental sensing algorithm is presented by combining the information between intensity image and that of laser structured light image. To see how effectively the algorithm applied to real environments, we developed a sensor system that can be mounted on a mobile robot and tested performance for a series of environments.

• International Journal of Reliability and Applications
• /
• 제16권2호
• /
• pp.55-79
• /
• 2015

It is well known that using any additional information in the estimation of unknown parameters with new sample of observations diminishes the sampling units needed and minimizes the risk of new estimators. There are many rational reasons to assure that the existence of additional information in practice and there exists many practical cases in which additional information is available in the form of target value (initial value) about the unknown parameters. This article is described the problem of how the prior initial value about the unknown parameters can be utilized and combined with classical Bayes estimator to get a new combination of Bayes estimator and prior value to improve the properties of the new combination. In this article, two classes of Bayes-shrinkage and preliminary test Bayes-shrinkage estimators are proposed for the scale parameter of exponential distribution. The bias, risk and risk ratio expressions are derived and studied. The performance of the proposed classes of estimators is studied for different choices of constants engaged in the estimators. The comparisons, conclusions and recommendations are demonstrated.

• Journal of information and communication convergence engineering
• /
• 제12권2호
• /
• pp.75-82
• /
• 2014

Dispersion management (DM), optical phase conjugation (OPC), and the combination of DM and OPC are promising techniques to compensate for optical signal distortion due to group velocity dispersion and nonlinear Kerr effects. The system performance improvement in DM links combined with OPC has been reported; however, the fixed residual dispersion per span (RDPS) usually used in these links restricts the flexibility of link configuration. Thus, in this paper, a flexible optical link configuration with artificially distributed single-mode fiber (SMF) lengths and RDPS in the combination of DM and OPC is proposed. Simulation results show that the best artificial distribution pattern is the gradually descending distribution of SMF lengths and the gradually ascending distribution of RDPS, as the number of fiber spans is increased, regardless of the average RDPS, the optimal net residual dispersion, and the dispersion coefficient of the dispersion compensating fiber.

• Journal of Information Technology Applications and Management
• /
• 제13권4호
• /
• pp.181-195
• /
• 2006

In recent years, increasing attention has been paid on how to manage a successful relationships between the receiver and provider of the Information Systems(IS) outsourcing services. This study explores the sources of influence for successful outsourcing relationships. Based on the Jaworski's control combination model for marketing, we (1) propose a management combination model of outsourcing relationships where trust is introduced as an intervening variable and outsourcing compleyity as a moderating variable, (2) test this model using a sample of 94 outsourcing projects in 36 organizations that have outsourced there IS functions to external service providers. The results indicated several significant findings. First, the deployment of management mechanism such as contract mechanism and trust building has a significant effect on the outsourcing success through trust only. Second, complexity of outsourcing activities has a moderating effect on relationship management combinations.

• Journal of information and communication convergence engineering
• /
• 제17권4호
• /
• pp.227-233
• /
• 2019

In long-haul optical communication system consisting of standard single-mode fiber spans and fiber amplifiers, such as the erbium-doped fiber amplifier, performance is deteriorated by signal distortion due to chromatic dispersion and nonlinearity of the fiber. A combination of dispersion management and optical phase conjugation is an effective technique to compensate for the distortion. In an optical link configured by this combination, a dispersion map mainly affects the compensation of the distorted optical signals. This paper proposes new dispersion maps configured by the decaying or expanding distribution of residual dispersion per span (RDPS) in a dispersion-managed link combined with a midway optical phase conjugator. The effect of the proposed dispersion maps on the compensation for distorted 24 channel × 40 Gbps wavelength-division multiplexed signals was assessed through numerical simulation. It was confirmed that all the proposed dispersion maps are most appropriate for the compensation and, furthermore, for the flexibility of link configuration than conventional links.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2014년도 추계학술발표대회
• /
• pp.693-694
• /
• 2014

This short paper briefly describes the proposed music recommendation method that provides suitable music pieces to a listener depending on both listeners' ratings and content of music pieces. The proposed method consists of two methods. First, listeners' ratings prediction method is a combination the traditional user-based and item-based collaborative filtering methods. Second, genre classification method is a combination of feature extraction and classification procedures. The feature extraction step obtains audio signal information and stores it in data structure, while the second one classifies the music pieces into various genres using decision tree algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제6권1호
• /
• pp.287-304
• /
• 2012

Spectrum sensing is an important functionality for cognitive users to look for spectrum holes before taking transmission in dynamic spectrum access model. Unlike previous works that assume perfect knowledge of the SNR of the signal received from the primary user, in this paper we consider a realistic case where the SNR of the primary user's signal is unknown to both fusion center and cognitive radio terminals. A Kalman filter based adaptive Takagi and Sugeno's fuzzy system is designed to make the global spectrum sensing decision based on the observed energies from cognitive users. With the capacity of adapting system parameters, the fusion center can make a global sensing decision reliably without any requirement of channel state information, prior knowledge and prior probabilities of the primary user's signal. Numerical results prove that the sensing performance of the proposed scheme outperforms the performance of the equal gain combination based scheme, and matches the performance of the optimal soft combination scheme.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권3호
• /
• pp.1722-1737
• /
• 2019

We developed an insider threat detection model to be used by organizations that repeat tasks at regular intervals. The model identifies the best combination of different feature selection algorithms, unsupervised learning algorithms, and standard scores. We derive a model specifically optimized for the organization by evaluating each combination in terms of accuracy, AUC (Area Under the Curve), and TPR (True Positive Rate). In order to validate this model, a four-year log was applied to the system handling sensitive information from public institutions. In the research target system, the user log was analyzed monthly based on the fact that the business process is processed at a cycle of one year, and the roles are determined for each person in charge. In order to classify the behavior of a user as abnormal, the standard scores of each organization were calculated and classified as abnormal when they exceeded certain thresholds. Using this method, we proposed an optimized model for the organization and verified it.