Gyoo-Seok Choi

Rough set approach for clustering categorical data using information-theoretic dependency measure

A variety of clustering algorithms exists to group objects having similar characteristics. But the implementations of many of those algorithms are challenging in the process of dealing with categorical data. While some of the algorithms cannot handle categorical data, others are unable to handle uncertainty within categorical data in nature. This is prerequisite for clustering categorical data which also deal with uncertainty. An algorithm, termed minimum-minimum roughness (MMR) was proposed, which uses the rough set theory in order to deal with the above problems in clustering categorical data. Later many algorithms has developed to improve the handling of hybrid data. This research proposes information-theoretic dependency roughness (ITDR), another technique for categorical data clustering taking into account information-theoretic attributes dependencies degree of categorical-valued information systems. In addition, it is second to none of all its predecessors; MMR, MMeR, SDR and standard-deviation of standard-deviation roughness (SSDR). Experimental results on two benchmark UCI datasets show that ITDR technique is better with the baseline categorical data clustering technique with respect to computational complexity and the purity of clusters.

A variable-precision information-entropy rough set approach for job searching

Data mining is the process of discovering hidden, non-trivial patterns in large amounts of data records in order to be used very effectively for analysis and forecasting. Because hundreds of variables give rise to a high level of redundancy and dimensionality with time complexity, they are more likely to have spurious relationships, and even the weakest relationships will be highly significant by any statistical test. Hence cluster analysis is a main task of data mining and is the task of grouping a set of objects in such a way that objects in the same group are more similar to each other than to those in other groups. In this paper system implementation is of great significance, which defines a new definition based on information-theoretic entropy and analyzes the analog behaviors of objects at hand so as to address the measurement of uncertainties in the classification of categorical data. The sources were taken from a survey aimed to identify of job guidance from students in high school at PyeongTaek. We show how variable precision information-entropy based rough set can be used to group students in each section. It is proved that the proposed method has the more exact classification than the conventional in attributes more than 10 and that is more effective in job searching for students.

Uncertainty Improvement of Incomplete Decision System using Bayesian Conditional Information Entropy

Abstract Based on the indiscernible relation of rough set, the inevitability of superposition and inconsistency of data makes the reduction of attributes very important in information system. Rough set has difficulty in the difference of attribute reduction between consistent and inconsistent information system. In this paper, we propose the new uncertainty measure and attribute reduction algorithm by Bayesian posterior probability for correlation analysis between condition and decision attributes. We compare the proposed method and the conditional information entropy to address the uncertainty of inconsistent information system. As the result, our method has more accuracy than conditional information entropy in dealing with uncertainty via mutual information of condition and decision attributes of information system. Keywords : Rough Set, Indiscernibility Relation, Conditional Information Entropy, Uncertainty, Bayesian Theory * 종신회원, 청운대학교 컴퓨터학과 ** 정회원, 중부대학교 컴퓨터학과 (교신저자)접수일자: 2014년 11월 11일, 수정일자: 2014년 12월 9일게재확정일자: 2014년 12월 12일 Received: 11 November, 2014 / Revised: 9 December, 2014Accepted: 12 December, 2014

Digital Modeling and Control of Multiple Time-Delayed Systems via SVD

Delays by controller-to-actuator and sensor-to-controller deteriorate control performance and could destabilize the overall system. In this paper, a new approximated discretization method and digital design for control systems with multiple delays is proposed. Based on a procedure for the generation of impulse response data, the multiple fractional/integer time-delayed continuous-time system is transformed to a discrete-time model with multiple integer time delays. To implement the digital modeling, the singular value decomposition (SVD) of a Hankel matrix together with a energy loss level is employed to obtain an extended discrete-time state space model. Then, the extended discrete-time state space model of the control system is reformulated as an integer time-delayed discrete-time system by computing its observable canonical form. The proposed method can closely approximate the step response of the original continuous time-delayed control system by choosing various of energy loss level. Illustrative example is simulated to demonstrate the effectiveness of the developed method.

Detecting user status from smartphone sensor data

Due to the high increment in usage and built-in advanced technology of smartphones, human activity recognition relying on smartphone sensor data has become a focused research area. In order to reduce noise of collected data, most of previous studies assume that smartphones are fixed at certain positions. This strategy is impractical for real life applications. To overcome this issue, we here investigate a framework that allows detecting the status of a traveller as idle or moving regardless the position and the direction of smartphones. The application of our work is to estimate the total energy consumption of a traveller during a trip. A number of experiments have been carried out to show the effectiveness of our framework when travellers are not only walking but also using primitive vehicles like motorbikes.

A study on Interactive-type Exhibition Using Fractal Images

Recent exhibitions paradigm is changing from the existing unidirectional oriented exhibition form to a form of interactive hands-on exhibits that viewers can get and realistically feel a variety of information. Hands-on exhibit embodies the human interface by utilizing light, sound, pressure, etc. in time and space. In this paper, we have studied the creation of fractal image by the Mandelbrot technique and proposed the interaction method for it to be converted into a variety of forms. By using the proposed method, a variety of image transformation such as printmaking effect, sketch effect, Pop Art effect can be performed, according to clicking a certain fraction on the created fractal image screen by a user mouse. Interactive image generated in this study are expected to be used for trade shows, promotional products, media art design.

Creation of Approximate Rules based on Posterior Probability

In this paper the patterns of information system is reduced so that control rules can guarantee fast response of queries in database. Generally an information system includes many kinds of necessary and unnecessary attribute. In particular, inconsistent information system is less likely to acquire the accuracy of response. Hence we are interested in the simple and understandable rules that can represent useful patterns by means of rough entropy and Bayesian posterior probability. We propose an algorithm which can reduce control rules to a minimum without inadequate patterns such that the implication between condition attributes and decision attributes is measured through the framework of rough entropy. Subsequently the validation of the proposed algorithm is showed through test information system of new employees appointment.

Using Bayesian posterior probability in confidence of attributes for feature selection

Rough set theory is an efficient reduction technique to deal with vagueness and uncertainty. Many studies have been accomplished for the feature selection while they have been carried out to trade off the sophisticated process of feature selection algorithm against the robustness and accuracy of reducts. In this paper, a new Bayesian posterior probability-based QuickReduct BPPQR measure is introduced to determine the optimal attributes with the accurate strength of the association among the indiscernible subsets. Therefore, a new rough entropy-based QuickReduct algorithm which focuses on the reduction of redundant attributes is proposed in order to extract the optimal reduct and the core. The performance of the system is evaluated in MATLAB on several benchmark datasets with resides in UCI machine learning repository. The proposed heuristic approach can cope with the drawbacks of the conventional one, and the satisfying performances have been carried out in the process of feature selection in decision systems.

A Study on Effective Information Transfer Technique between Personal Computer and Portable Equipment

In this paper, the effective information transfer technique between computing device and portable device that is connected through USB interface has been studied. The proposed technique is realized by the method which portable device is perceived into optical media such as CD-ROM or CD-R/W through SCSI commands on USB interface. Through the proposed system in this study, we maximized user convenience by executing diverse works using stored data in a portable device without installing extra USB driver corresponding to portable device, when user data is transmitted between computing device and portable device. We conducted experimental tests to verify the performance of the proposed system through implementing of test system. As a result, we confirmed the technique is comparatively superior to existing method in transmitting speed and user convenience.

The Case Study of Cancer Diagnosis System Based on Kernel Method and Genetic Approach

One of the most important problems in bioinformatics is how to extract the useful information from a huge amount of data, and make decision in diagnosis, prognosis, and medical treatment applications. To deal with this problem, many approaches have been proposed. The main goal of our study is to figure out an efficient method to achieve a cancer diagnosis system with high accuracies, and good adaptability to various types of data set. For accomplishing this goal, we proposed a new kernel function which is defined as the weighted sum of a set of different types of basic kernel functions and also its learning method based on GA. The experimental results on clinical datasets are proved that our proposed approach with the novel kernel function and its learning method has a higher prediction rate, comparable and sometimes better performance than the previous ones.