• ETRI Journal
• /
• v.41 no.2
• /
• pp.235-241
• /
• 2019

In this paper, we propose a rank-weighted reconstruction feature to improve the robustness of a feed-forward deep neural network (FFDNN)-based acoustic model. In the FFDNN-based acoustic model, an input feature is constructed by vectorizing a submatrix that is created by slicing the feature vectors of frames within a context window. In this type of feature construction, the appropriate context window size is important because it determines the amount of trivial or discriminative information, such as redundancy, or temporal context of the input features. However, we ascertained whether a single parameter is sufficiently able to control the quantity of information. Therefore, we investigated the input feature construction from the perspectives of rank and nullity, and proposed a rank-weighted reconstruction feature herein, that allows for the retention of speech information components and the reduction in trivial components. The proposed method was evaluated in the TIMIT phone recognition and Wall Street Journal (WSJ) domains. The proposed method reduced the phone error rate of the TIMIT domain from 18.4% to 18.0%, and the word error rate of the WSJ domain from 4.70% to 4.43%.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.29 no.5
• /
• pp.1007-1017
• /
• 2019

Currently, most biometrics system authenticates users by using single biometric information. This method has many problems such as noise problem, sensitivity to data, spoofing, a limitation of recognition rate. One method to solve this problems is to use multi biometric information. The multi biometric authentication system performs information fusion for each biometric information to generate new information, and then uses the new information to authenticate the user. Among information fusion methods, a score-level fusion method is widely used. However, there is a problem that a normalization operation is required, and even if data is same, the recognition rate varies depending on the normalization method. A rank-level fusion method that does not require normalization is proposed. However, a existing rank-level fusion methods have lower recognition rate than score-level fusion methods. To solve this problem, we propose a rank-level fusion method with higher recognition rate than a score-level fusion method using correlation coefficient. The experiment compares recognition rate of a existing rank-level fusion methods with the recognition rate of proposed method using iris information(CASIA V3) and face information(FERET V1). We also compare with score-level fusion methods. As a result, the recognition rate improve from about 0.3% to 3.3%.

• Journal of Information Technology Services
• /
• v.22 no.1
• /
• pp.57-74
• /
• 2023

It has been pointed out as a limitation that the rank of some efficient DMUs(decision making units) cannot be discriminated due to the relativity nature of efficiency measured by DEA(data envelopment analysis), comparing the production structure. Recently, to solve this problem, a DEA-SNA(social network analysis) model that combines SNA techniques with data envelopment analysis has been studied intensively. Several models have been proposed using techniques such as eigenvector centrality, pagerank centrality, and hypertext induced topic selection(HITS) algorithm, but DMUs that cannot be ranked still remain. Moreover, in the process of extracting latent information within the DMU group to build effective network, a problem that violates the basic assumptions of the DEA also arises. This study is meaningful in finding the cause of the limitations by comparing and analyzing the characteristics of the DEA-SNA model proposed so far, and based on this, suggesting the direction and possibility to develop more advanced model. Through the results of this study, it will be enable to further expand the field of research related to DEA.

• The Journal of Korean Physical Therapy
• /
• v.32 no.6
• /
• pp.394-399
• /
• 2020

Purpose: The aim of this study was to determine the effects of lower rib cage lateral expansion limitation on the maximal inspiratory and expiratory pressures and on abdominal muscle activity during maximal respiratory breathing in healthy subjects. Methods: Fifteen healthy male subjects voluntarily participated in this cross-sectional study. During maximal breathing, maximal inspiratory and expiratory pressures were measured, and abdominal muscle activity was determined with using surface electromyography. Also, the measurement was repeated with using a non-elastic belt to the lower rib cage for limiting of lateral expansion. A Wilcoxon signed-rank test was performed for obtaining the statistical difference with a significance level of 0.05. Results: The findings of this study are as follows: 1) There were no significant differences in maximal inspiratory and expiratory pressure with and without lower rib cage lateral expansion (p>0.05), 2) There was no significant difference in abdominal muscle activity during the maximal inspiratory phase (p>0.05). However, right external oblique muscle activity decreased significantly during maximum exhalation with lower rib expansion limitation (p<0.05). Conclusion: The results of the current study indicate that a non-elastic belt was effective in decreasing right external oblique muscle activity during forced expiratory breathing in healthy subjects.

• Journal of KIISE:Databases
• /
• v.37 no.6
• /
• pp.343-350
• /
• 2010

We can get answers we want to know via questioning in Knowledge Search Service (KSS) based on Q&A Community. However, it is getting more difficult to find credible documents in enormous documents, since many anonymous users regardless of credibility are participate in answering on the question. In previous works in KSS, researchers evaluated the quality of documents based on textual information, e.g. recommendation count, click count and non-textual information, e.g. answer length, attached data, conjunction count. Then, the evaluation results are used for enhancing search performance. However, the non-textual information has a problem that it is difficult to get enough information by users in the early stage of Q&A. The textual information also has a limitation for evaluating quality because of judgement by partial factors such as answer length, conjunction counts. In this paper, we propose the QualityRank algorithm to improve the problem by textual and non-textual information. This algorithm ranks the relevant and credible answers by considering textual/non-textual information and user centrality based on Social Network Analysis(SNA). Based on experimental validation we can confirm that the results by our algorithm is improved than those of textual/non-textual in terms of ranking performance.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.6
• /
• pp.737-744
• /
• 2004

Retrieving relevant information well corresponding to the user`s request from web is a crucial task of search engines. However, most of conventional search engines based on pattern matching schemes to queries have a limitation that is not easy to provide results corresponding to the user`s request due to the uncertainty of queries. To overcome the limitation in this paper, we propose a framework for knowledge-based semantic meta-search engines with the following five processes: (i) Query formation, (ii) Query expansion, (iii) Searching, (iv) Ranking recreation, and (v) Knowledge base. From simulation results on english-based web documents, we can see that the Proposed knowledge-based semantic meta-search engine provides more correct and better searching results than those obtained by using the Google.

• Journal of Korea Technology Innovation Society
• /
• v.11 no.3
• /
• pp.376-399
• /
• 2008

Science and technology (S&T) is one of the most important elements in a nation's competitiveness. In an effort to strengthen their national competitiveness, all countries are focusing on upgrading the level of eir S&T. With these factors in mind, Korea has increased its support of national research and development (R&D). In recent years, this added support has resulted in an increased interest in the effectiveness of R&D. We have made continuous efforts to enhance the accountability and effectiveness of R&D by strengthening performance evaluation and considering R&D evaluation results during the budget review (appropriation) process. In order to change to a performance based system, we need to develop objective and scientific indicators to measure and evaluate the quality of the research performance of R&D programs. One of the primary research outcomes is publications. The impact factor of publications is widely used to evaluate overall journal quality and the quality of the papers published therein. However, the use of impact factors has been criticised because they can vary greatly when works from different subject areas are compared. In order to overcome this limitation, we have developed three kinds of qualitative indicators, which are functions of the impact factor. Two of these qualitative indicators, Modified Rank Normalized Impact Factor and Ordinal Rank Normalized Impact Factor, are based on order statistics (rank) for all journals from a specific specialty. The third qualitative indicator, Relative Field Impact Factor, uses the average impact factor of all journals within a subject category. We also suggest a quantitative indicator, Percentage of Contribution. In this study, we suggest 4 indicators and use them to evaluate the performance of outcomes from three R&D programs supported by the Ministry of Education, Science & Technology. We also perform a simulation study to verify the effectiveness of the proposed indicators. It can be shown that the proposed Ordinal Rank Normalized Impact Factor is the most reliable and effective indicator for comparing research performance across subject categories. However, we recommend using previous indicators in combination with the proposed indicators in this study for the research evaluation of R&D programs.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.13 no.1
• /
• pp.211-219
• /
• 2012

This study was performed to search for correlation between activity limitation level and participation restriction of children with cerebral palsy by investigating the activity limitation level through the use of the GMFCS E&R and the MACS and by figuring out participation restriction level through questionnaire survey. This study was performed, from May 1, 2010 to August 31, 2010, on 152 children with cerebral palsy ranging from 4 to 12 years, who are receiving rehabilitation therapy in the hospitals and community clinics in Busan and Gyeongnam province. The levels of activity limitation were assessed by using GMFCS E&R and MACS, and a questionnaire survey was conducted for participation restriction of the participating children. Spearman rank correlation was used for correlation analysis with the statistical software, SPSS 12.0. Majority of the children scored level 5 in both GMFCS E&R and MACS which is high frequency of activity limitation levels. Children in these levels also showed severe or complete participation restriction for mobility, education, and social relations. The analysis showed a significantly positive correlation between activity limitation level and participation restriction. Therefore this study may be useful in assessing the functional movement component of participation in children with cerebral palsy and developing the intervention plan for participation.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.64 no.4
• /
• pp.333-336
• /
• 2015

Recently, the power load is growing larger and because of the environmental limitation of generation, the expansion of generation facilities are becoming more difficult. For that reason the importance of the demand-side resources come to be higher. One method of the demand-side resource, the DLC Program, has executed, and moreover, the loads which are available to be controlled are increasing. It should be considered of some kinds of power system components such as DLCs, because the fact that using the demand resources will be an important part of the power system. This paper considers the power factor of the load-bus which is shedded in the direct load control program. and then analyze the power system using flow sensitivity and voltage sensitivity. In this paper, we assumed two scenarios through the rank of the load power factor at each bus and to compare and evaluate each case, we used Power World for the simulation.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.24 no.1
• /
• pp.42-54
• /
• 1996

Many of urban parks which have been designated for urban planning regulation are not well developed because the priority of investment is inferior to that of other public facilities. Park development is an urgent urban policy and it is hard to decide priority among parks. Therefore, it is necessary to evaluate the priority of park development. Thus the Analytic Hierarchy Process has been introduced because it is suitable for achieving objectivity by collecting the professional opinions. The goals of park development are evaluated by AHP method and the priorities among goals are conservation of natural environment, satisfying the residents' residents' needs, population density, regional pollution, land acquisition cost, ratio of greeneries and facilities within park areas. Priority of park development is decided by the standardization of each individual inventory data and considering of the professional opinions. Among the multi-attribute judgement methods, AHP method may be the most widely used. Despite the critical evaluations of the AHP, such as limitation of the nine-point scale and the rank reversal problem, application of AHP to decide the priority of park development is suitable for actual circumstances.