• Journal of Ubiquitous Convergence Technology
• /
• v.2 no.1
• /
• pp.40-50
• /
• 2008

The problem of assigning customers to satellite channels is a difficult combinatorial optimization problem and is NP-complete. For this combinatorial optimization problem, standard optimization methods take a large computation time and so genetic algorithms (GA) and ant colony optimization (ACO) can be used to obtain the best and/or optimal assignment of customers to satellite channels. In this paper, we present a comparative study of GA and ACO to this problem. Various issues related to genetic algorithms approach to this problem, such as solution representation, selection methods, genetic operators and repair of invalid solutions are presented. We also discuss an ACO for this problem. In ACO methodology, three strategies, ACO with only ranking, ACO with only max-min ant system (MMAS), and ACO with both ranking and MMAS, are considered. A comparison of these two approaches (i,e., GA and ACO) with the standard optimization method is presented to show the advantages of these approaches in terms of computation time.

• Communications for Statistical Applications and Methods
• /
• v.16 no.6
• /
• pp.1005-1012
• /
• 2009

The area under the curve(AUC) is commonly used as a measure of the receiver operating characteristic(ROC) curve which displays the performance of a set of binary classifiers for all feasible ratios of the costs associated with true positive rate(TPR) and false positive rate(FPR). In the bipartite ranking problem where one has to compare two different observations and decide which one is "better", the AUC measures the quantity that ranking score of a randomly chosen sample in one class is larger than that of a randomly chosen sample in the other class and hence, the function which maximizes an AUC of bipartite ranking problem is different to the function which maximizes (minimizes) accuracy (misclassification error rate) of binary classification problem. In this paper, we develop a way to construct the unified framework for AUC maximizer including support vector machines based on maximizing large margin and logistic regression based on estimating posterior probability. Moreover, we develop an efficient algorithm for the proposed unified framework. Numerical results show that the propose unified framework can treat various methodologies successfully.

• The Korean Journal of Applied Statistics
• /
• v.32 no.5
• /
• pp.763-782
• /
• 2019

This paper proposes a probabilistic document ranking model incorporating term dependencies. Document ranking is a fundamental information retrieval task. The task is to sort documents in a collection according to the relevance to the user query (Qin et al., Information Retrieval Journal, 13, 346-374, 2010). A probabilistic model is a model for computing the conditional probability of the relevance of each document given query. Most of the widely used models assume the term independence because it is challenging to compute the joint probabilities of multiple terms. Words in natural language texts are obviously highly correlated. In this paper, we assume a multinomial distribution model to calculate the relevance probability of a document by considering the dependency structure of words, and propose an information retrieval model to rank a document by estimating the probability with the maximum entropy method. The results of the ranking simulation experiment in various multinomial situations show better retrieval results than a model that assumes the independence of words. The results of document ranking experiments using real-world datasets LETOR OHSUMED also show better retrieval results.

• Journal of Korea Multimedia Society
• /
• v.13 no.8
• /
• pp.1245-1255
• /
• 2010

In current Web 2.0 environment, one of the most core technology is social bookmarking which users put tags and bookmarks to their interesting Web pages. The main purpose of social bookmarking is an effective information service by use of retrieval, grouping and share based on user's bookmark information and tagging result of their interesting Web pages. But, current social bookmarking system uses the number of bookmarks and tag information separately in information retrieval, where the number of bookmarks stand for user's degree of interest on Web contents, information retrieval, and classification serve the purpose of tag information. Because of above reason, social bookmarking system does not utilize effectively the bookmark information and tagging result. This paper proposes a Web contents ranking algorithm combining bookmarks and tag information, based on preceding research on associative tag extraction by tag clustering. Moreover, we conduct a performance evaluation comparing with existing retrieval methodology for efficiency analysis of our proposed algorithm. As the result, social bookmarking system utilizing bookmark with tag, key point of our research, deduces a effective retrieval results compare with existing systems.

• The Korean Journal of Applied Statistics
• /
• v.30 no.6
• /
• pp.1015-1025
• /
• 2017

Bayesian personalized ranking (BPR) is a state-of-the-art recommendation system techniques for implicit feedback data. Unfortunately, there might be a loss of information because the BPR model considers only the binary transformation of implicit feedback that is non-binary data in most cases. We propose a modified BPR method using a level of confidence based on the size or strength of implicit feedback to overcome this limitation. The proposed method is useful because it still has a structure of interpretable models for underlying personalized ranking i.e., personal pairwise preferences as in the BPR and that it is capable to reflect a numerical size or the strength of implicit feedback. We propose a computation algorithm based on stochastic gradient descent for the numerical implementation of our proposal. Furthermore, we also show the usefulness of our proposed method compared to ordinary BPR via an analysis of steam video games data.

• Journal of the Korean Institute of Gas
• /
• v.7 no.1 s.18
• /
• pp.1-9
• /
• 2003

A Multi-distinction Equipment Screening Algorithm (MESA) is proposed. It selectively integrates Dow's F&EI as its process hazard index technique and ESA (Equipment Screening Algorithm) as qualitative hazard classification technique, and retrieves a detailed list of hazardous equipments with the total hazard indices of those equipments. The inherent expert system, which includes the accident scenarios of the equipments and processes and experts' views of them, narrows further down the list of hazardous equipments and recommends only the most notable candidates. Through the case study of distinguishing the hazardous ranking of the equipments of the LPG underground storage process, using the expert system or not, the applicability of MESA has been validated. Taking the characteristics of the process equipments with hazardous ranking in the point of process intrinsic safety, this proposed algorithm would contribute to providing engineers or managers with information on constructing safely devices and mitigation devices and on scheduling emergency response planning.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.1
• /
• pp.7-15
• /
• 2014

Image retrieval has become a huge part of computer vision and data mining. Although commercial image retrieval systems such as Google show great performances, the improvement on the performances are constantly on demand because of the rapid growth of data on web space. To satisfy the demand, many re-ranking methods, which enhance the performances by reordering retrieved results with independent algorithms, has been proposed. Conventional re-ranking algorithms are based on the assumption that visual patterns are not used on initial image retrieval stage. However, image search engines in present have begun to use the visual and the assumption is required to be reconsidered. Also, though it is possible to suspect that integration of multiple retrieval systems can improve the overall performance, the research on the topic has not been done sufficiently. In this paper, we made the condition that other manner than cooperation cannot improve the ranking result. We evaluate the algorithm on toy model and show that propose module can improve the retrieval results.

• The KIPS Transactions:PartD
• /
• v.9D no.5
• /
• pp.745-754
• /
• 2002

The Google search site (http://www.google.com), which was introduced in 1998, implemented the PageRank algorithm for the first time. PageRank is a ranking method based on the link structure of the Web pages. Even though PageRank has been implemented and being used in various commercial search engines, implementation details did not get documented well, primarily due to business reasons. Implementation techniques introduced in [4,8] are not sufficient to produce PageRank values of Web pages. This paper explains the techniques[4,8], and suggests major data structure and four implementation techniques in order to apply the PageRank algorithm. The paper helps understand the methods of applying PageRank algorithm by means of showing a real system that produces PageRank values of Web pages.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.22 no.8
• /
• pp.96-103
• /
• 2008

In this paper, a new algorithm based on sensitivity method for alleviating overloads in power networks is presented to find the switching branches effectively. By applying the new sensitivity of the line flow with respect to the change of the branch impedance, both on and off switchings for alleviating overloads in power networks are performed systematically at once and an effective scheme for drawing up a plan to alleviate overloads of the lines is presented through screening a large number of line switching cases easily. The ranking of switching branches is calculated according to the new switching algorithm based on sensitivity method and the switching of the ranked branches is performed in the order of ranking until overloads are eliminated. In order to show the effects of this algorithm, it is applied to a small scale power system of IEEE 39-bus test system.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.32 no.10
• /
• pp.360-366
• /
• 1983

In this paper, operation cost of the system is calculated by the probabilistic simulation method. And it is proved that only 20 iterative simulations are enough to get the result obtain by the Monte Carlo simulation method which requires more than 1000 iterative simulations. In the probabilistic simulation method we use the ranking of line contingency which is derived from the line countingency selection algorithm proposed in (2). In using this method the nature of the sparsity of the power system is used.