• Journal of Digital Convergence
• /
• v.12 no.12
• /
• pp.365-370
• /
• 2014

The existing system consumes a considerable amount time and cost for extracting the patterns from whole sequences or misaligned sequences. In this paper, We propose the classification system, which creates the partition sequence sections using multiple sequence alignment method and extracts the sequential patterns from these section. These extracted patterns are accumulated motif candidate sets and then used the training sets of SVM classifier. This proposed system predicts a HPV-type(high/low) using the learned knowledges from known/unknown protein sequences and shows more improved precision, recall than previous system in 30% minimum support.

• Journal of information and communication convergence engineering
• /
• v.12 no.1
• /
• pp.8-13
• /
• 2014

We consider the sensor selection problem in large sensor networks where the goal is to find the best set of sensors that maximizes application objectives. Since sensor selection typically involves a large number of sensors, a low complexity should be maintained for practical applications. We propose a geometry-based sensor selection algorithm that utilizes only the information of sensor locations. In particular, by observing that sensors clustered together tend to have redundant information, we theorize that the redundancy is inversely proportional to the distance between sensors and seek to minimize this redundancy by searching for a set of sensors with the maximum average distance. To further reduce the computational complexity, we perform an iterative sequential search without losing optimality. We apply the proposed algorithm to an acoustic sensor network for source localization, and demonstrate using simulations that the proposed algorithm yields significant improvements in the localization performance with respect to the randomly generated sets of sensors.

• Communications for Statistical Applications and Methods
• /
• v.6 no.2
• /
• pp.543-564
• /
• 1999

The principal aim of a sequential phase I clinical trial in which the toxicity reponses of a group of patient(s) determine the dose level of the next patient(s) group is to estimate the maximal tolerated dose(MTD) of a new drug, In this paper we compared with a simulation study the performance of the MTD estimates that are determined by a stopping rule in a design and also those that are determined by analyzing the data after a clinical trial is terminated. To the latter belong the mean median mode and maximum likelihood estimates. For the Standard Methods the stopping rule MTD is quite inefficient but the median MTD has a best efficiency and is robust with respect to the three different toxicity curves. The problem of non-convergence of MLE MTD is severe. A more improved MTD estimate is produced by combining the advantages of the various MTD estimates and its efficiency is better than the single median MTD estimate especially for the toxicity curve of an unlucky choice of dose levels. The simulation results suggest that simple types of phase I designs can be combined with relatively standard analytic techniques to provide a more efficient MTD estimate.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.33A no.5
• /
• pp.1-9
• /
• 1996

In this paper, we propose an efficient computational algorithm that can reduce significantly the computational complexity of the ML optimum multiuser detector known as the most excellent detector in synchronous code division multiple access channels. The proposed detector uses the sequential detection algorithm based on the alternating maximization appraoch to obtain the ML estimates. As initial estimates for this sequential algorithm, we can use the estimated values obtained by the conventional single-user detector, the linear decorrelating multiuser detector, or the decorrelating decision-feedback muliuser detector, the linear decorrelating multiuser detector, or the decorrelating decision-feedback multiuser detector. We have performed computer simulations in order to see the convergence behaviors and the detection performance of the propsoed algorithm in terms of initial algorithms and the number of users, and then to compare the computational complexity with that of the ML optimum multiuser detector. From the results, we have seen that the proposed alternating maximization detector has nearly equal detction performance with that of the ML optimum multiuser detctor in only a few iteration.

• East Asian mathematical journal
• /
• v.28 no.3
• /
• pp.293-303
• /
• 2012

Sequential quadratic programming (SQP) has been one of the most important methods for solving nonlinear constrained optimization problems. Recently, filter method, proposed by Fletcher and Leyffer, has been extensively applied for its promising numerical results. In this paper, we present and study an active set SQP-filter algorithm for inequality constrained optimization. The active set technique reduces the size of quadratic programming (QP) subproblem. While by the filter method, there is no penalty parameter estimate. Moreover, Maratos effect can be overcome by filter technique. Global convergence property of the proposed algorithm are established under suitable conditions. Some numerical results are reported in this paper.

• Journal of Multimedia Information System
• /
• v.8 no.1
• /
• pp.31-34
• /
• 2021

The classical HMM is defined by a parameter triple �� = (��, A, B), where each parameter represents a collection of probability distributions: initial state, state transition and output distributions in order. This paper proposes a new stationary parameter e = (e1, e2, …, eN) where N is the number of states and et = P(|xt = i, y) for describing how an input pattern y ends in state xt = i at time t followed by nothing. It is often said that all is well that ends well. We argue here that all should end well. The paper sets the framework for the theory and presents an efficient inference and training algorithms based on dynamic programming and expectation-maximization. The proposed model is applicable to analyzing any sequential data with two or more finite segmental patterns are concatenated, each forming a context to its neighbors. Experiments on online Hangul handwriting characters have proven the effect of the proposed augmentation in terms of highly intuitive segmentation as well as recognition performance and 13.2% error rate reduction.

• International journal of advanced smart convergence
• /
• v.11 no.3
• /
• pp.23-27
• /
• 2022

In this paper, we explore a new type of Android evasive malware based on the Sequential Probability Ratio Test (SPRT) that does not perform malicious task when it discerns that dynamic analyzer is input generator. More specifically, a new type of Android evasive malware leverages the intuition that dynamic analyzer provides as many inputs within a certain amount of time as possible to Android apps to be tested, while human users generally provide necessary inputs to Android apps to be used. Under this intuition, it harnesses the SPRT to discern whether dynamic analyzer runs in Android system or not in such a way that the number of inputs per time slot exceeding a preset threshold is regarded as evidence that inputs are provided by dynamic analyzer, expediting the SPRT to decide that dynamic analyzer operates in Android system and evasive malware does not carry out malicious task.

• Journal of the Korea Convergence Society
• /
• v.8 no.8
• /
• pp.71-75
• /
• 2017

Due to the spread of SNS, web-based consumer-generated data is increasing exponentially. It is important in many fields to accurately extract what is appropriate for the user's interest in a large amount of data. It is especially important for business mangers to establish marketing policies to find the right customers for them in many users. In this paper, we try to obtain important information centering on customers who are interested in each account through Twitter follow - following relationship. Because Twitter's current follower relationships do not reflect the user's interests, we try to figure out the details of interest using keyword extraction methods for tweets of followers. To do this, we select two domestic commercial Twitter accounts and apply the sequential pattern evaluation index to the mining key phrase of the text data collected from the follower.

• Journal of Korea Multimedia Society
• /
• v.6 no.5
• /
• pp.788-796
• /
• 2003

This paper proposes using dynamic momentum for squential learning method. Using The dynamic momentum improves convergence speed and performance by the variable momentum, also can identify it in the RMSE(root mean squared error). The proposed method is reflected using variable momentum according to current state. While static momentum is equally influenced on the whole, dynamic momentum algorithm can control the convergence rate and performance. According to the variable change of momentum by training. Unlike former classification and regression problems, this paper confirms both performance and regression rate of the dynamic momentum. Using RMSE(root mean square error ), which is one of the regression methods. The proposed dynamic momentum has been applied to the kernel adatron and kernel relaxation as the new sequential learning method of support vector machine presented recently. In order to show the efficiency of the proposed algorithm, SONAR data, the neural network classifier standard evaluation data, are used. The simulation result using the dynamic momentum has a better convergence rate, performance and RMSE than those using the static moment, respectively.

• Journal of Information Display
• /
• v.3 no.2
• /
• pp.19-25
• /
• 2002

In this paper, a new target extraction algorithm is proposed, in which the coordinates of target are obtained adaptively by using the difference image information and the optical BPEJTC(binary phase extraction joint transform correlator) with which the target object can be segmented from the input image and background noises are removed in the stereo vision system. First, the proposed algorithm extracts the target object by removing the background noises through the difference image information of the sequential left images and then controlls the pan/tilt and convergence angle of the stereo camera by using the coordinates of the target position obtained from the optical BPEJTC between the extracted target image and the input image. From some experimental results, it is found that the proposed algorithm can extract the target object from the input image with background noises and then, effectively track the target object in real time. Finally, a possibility of implementation of the adaptive stereo object tracking system by using the proposed algorithm is also suggested.