• Communications for Statistical Applications and Methods
• /
• 제12권3호
• /
• pp.659-672
• /
• 2005

This paper evaluates discretization of continuous variables to select relevant variables for supervised learning using mutual information. Three discretization methods, MDL, Histogram and 4-Intervals are considered. The process of discretization and variable subset selection is evaluated according to the classification accuracies with the 6 real data sets of UCI databases. Results show that 4-Interval discretization method based on quantiles, is robust and efficient for variable selection process. We also visually evaluate the appropriateness of the selected subset of variables.

• 대한산업공학회지
• /
• 제27권4호
• /
• pp.394-405
• /
• 2001

Since most real-world application data involve continuous-valued attributes, properly addressing the discretization process for constructing a decision tree is an important problem. A continuous-valued attribute is typically discretized during decision tree generation by partitioning its range into two intervals recursively. In this paper, by removing the restriction to the binary discretization, we present a hybrid multi-interval discretization algorithm for discretizing the range of continuous-valued attribute into multiple intervals. On the basis of experiment using semiconductor etching machine, it has been verified that our discretization algorithm constructs a more efficient incremental decision tree compared to previously proposed discretization algorithms.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권2호
• /
• pp.374-388
• /
• 2011

Biometric discretization is a process of transforming continuous biometric features of an identity into a binary bit string. This paper mainly focuses on improving the global discretization method - a discretization method that does not base on information specific to each user in bitstring extraction, which appears to be important in applications that prioritize strong security provision and strong privacy protection. In particular, we demonstrate how the actual performance of a global discretization could further be improved by embedding a global discriminative feature selection method and a Linearly Separable Subcode-based encoding technique. In addition, we examine a number of discriminative feature selection measures that can reliably be used for such discretization. Lastly, encouraging empirical results vindicate the feasibility of our approach.

• 한국CDE학회논문집
• /
• 제11권3호
• /
• pp.223-233
• /
• 2006

Layered manufacturing(LM) is emerging as a new technology that enables the fabrication of three dimensional heterogeneous objects such as Multi-materials and Functionally Gradient Materials (FGMs). Among various types of heterogeneous objects, more attention has recently paid on the fabrication of FGMs because of their potentials in engineering applications. The necessary steps for LM fabrication of FGMs include representation and process planning of material information inside an FGM. This paper introduces a new process planning algorithm that takes into account the processing of material information. The detailed tasks are discretization (i.e., decomposition-based approximation of volume fraction), orientation (build direction selection), and adaptive slicing of heterogeneous objects. In particular, this paper focuses on the discretization process that converts all of the material information inside an FGM into material features like geometric features. It is thus possible to choose an optimal build direction among various pre-selected ones by approximately estimating build time. This is because total build time depends on the complexity of features. This discretization process also allows adaptive slicing of heterogeneous objects to minimize surface finish and material composition error. In addition, tool path planning can be simplified into fill pattern generation. Specific examples are shown to illustrate the overall procedure.

• Communications for Statistical Applications and Methods
• /
• 제18권1호
• /
• pp.89-102
• /
• 2011

이산화는 데이터 마이닝을 위한 전처리 과정으로서 연속형 변수를 이산형 변수로 바꾸는 과정이고, 이산화 시킨 데이터가 원래 가지고 있던 정보손실을 최소로 하면서 높은 분류정확도를 가지는 것을 목적으로 한다. 지금까지 많은 이산화 알고리즘이 제안되었는데, 본 논문에서는 분할 이산화와 병합 이산화의 관점에서 최근까지 제안된 대표적인 이산화 알고리즘들을 비교하고, 이산화 알고리즘이 가지고 있는 특성을 연구하였다. 또한 비교 연구한 이산화 알고리즘을 R코드로 작성하여 다른 연구에 사용할 수 있도록 하였다.

• 대한산업공학회지
• /
• 제17권1호
• /
• pp.117-126
• /
• 1991

When {X(t), t ${\geq}$ 0} is a Markov process representing time-varying system states, we develop efficient bounding methods for some time-dependent performance measures. We use the discretization technique for stochastically monotone Markov processes and a combination of discretization and uniformization for Markov processes with the stochastic convexity(concavity) property. Sufficient conditions for stochastic monotonocity and stochastic convexity of a Markov process are also mentioned. A simple example is given to demonstrate the validity of the bounding methods.

• 산업경영시스템학회지
• /
• 제28권1호
• /
• pp.1-7
• /
• 2005

Data mining is widely used for turning huge amounts of data into useful information and knowledge in the information industry in recent years. When analyzing data set with continuous values in order to gain knowledge utilizing data mining, we often undergo a process called discretization, which divides the attribute's value into intervals. Such intervals from new values for the attribute allow to reduce the size of the data set. In addition, discretization based on rough set theory has the advantage of being easily applied. In this paper, we suggest a discretization algorithm based on Rough Set theory and SOM(Self-Organizing Map) as a means of extracting valuable information from large data set, which can be employed even in the case where there lacks of professional knowledge for the field.

• 대한전기학회논문지
• /
• 제40권9호
• /
• pp.920-928
• /
• 1991

This paper proposes a design procedure based on the LQG/LTR (Linear Quadratic Gaussian/ Loop Transfer Recovery) method for a launch vehicle. Continuous-discrete type LQG/LTR compensators are designed using the e-transformation to overcome numerical problems occurring in the process of discretization. The e-LQG/LTR compensator using the e-transformation is compared width the z-LQG/LTR compensator using the z-transformation. The performance of the overall system controlled by the compensator is evaluated via simulations, which show that the discretization error problem is resolved and the control performances are satisfactory in the proposed compensator.

• 대한기계학회논문집
• /
• 제19권9호
• /
• pp.2165-2172
• /
• 1995

Optimization of mesh discretization has been proposed to improve the accuracy of limit analysis solution of collapse load by using the Rigid Body Spring Model(R. B. S. M) under the plane strain condition. Moreover, the fracture behaviour of materials was investigated by employing the fracture mechanism of a spring connecting the triangular rigid body element. It has been clarified that the collapse load and the geometry of slip boundary for optimized mesh discretization were close to those of the slip line solution. Further, the wedge-shaped fracture of a cylinder under a lateral load and the central fracture of a strip in the drawing process were well simulated.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제15권1호
• /
• pp.57-61
• /
• 2009

항목 기반의 순차 패턴 마이닝 기법들을 시공간 데이타에 적용하기 위해서는 시공간 속성 값에 대한 적절한 이산화가 필수적이다. 본 논문에서는 입력 데이타의 시공간적 상판 정보를 유지함과 동시에 데이타 수를 축소시킴으로써 마이닝 프로세스의 효율성을 높이는 이산화 기법을 제안한다. 제안된 기법은 선 단순화를 사용하여 궤적에 대한 근사치를 구함으로써 마이넘 단계에서 처리할 데이터 크기를 축소시킨다. 또한 단순화 된 궤적을 유사한 시공간적 특성을 가지는 논리적 그룹으로 군집화하여 데이터의 분포를 고려한 이산화를 수행한다. 실험을 통해 제안된 기법이 마이넝 프로세스의 효율성을 높일 뿐 아니라 보다 직관적이고 해석이 용이한 패턴을 도출하는 것을 보였다.