• The Journal of the Acoustical Society of Korea
• /
• v.43 no.2
• /
• pp.200-206
• /
• 2024

Computing music similarity is an indispensable component in developing music search service. This paper proposes a relevance weight of each chromagram vector for cover song identification in computing a music similarity function in order to boost identification accuracy. We derive a music similarity function using the relevance weight based on the probabilistic relevance model, where higher relevance weights are assigned to less frequently-occurring discriminant chromagram vectors while lower weights to more frequently-occurring ones. Experimental results performed on two cover music datasets show that the proposed music similarity improves the cover song identification performance.

• Journal of KIISE:Software and Applications
• /
• v.31 no.10
• /
• pp.1374-1385
• /
• 2004

The relevance feedback process uses information obtained from a user about an initially retrieved set of documents to improve subsequent search formulations and retrieval performance. In the extended Boolean model, the relevance feedback Implies not only that new query terms must be identified, but also that the terms must be connected with the Boolean AND/OR operators properly Salton et al. proposed a relevance feedback method for the extended Boolean model, called the DNF (disjunctive normal form) method. However, this method has a critical problem in generating a reformulated queries. In this study, we investigate the problem of the DNF method and propose a relevance feedback method using hierarchical clustering techniques to solve the problem. We show the results of experiments which are performed on two data sets: the DOE collection in TREC 1 and the Web TREC 10 collection.

• Journal of Korea Multimedia Society
• /
• v.24 no.8
• /
• pp.1122-1137
• /
• 2021

PM_2.5 concentration in Seoul could be predicted by deep neural network model. In this paper, the contribution of input factors to the model's prediction results is analyzed using the LRP(Layer-wise Relevance Propagation) technique. LRP analysis is performed by dividing the input data by time and PM concentration, respectively. As a result of the analysis by time, the contribution of the measurement factors is high in the forecast for the day, and those of the forecast factors are high in the forecast for the tomorrow and the day after tomorrow. In the case of the PM concentration analysis, the contribution of the weather factors is high in the low-concentration pattern, and that of the air quality factors is high in the high-concentration pattern. In addition, the date and the temperature factors contribute significantly regardless of time and concentration.

• The Mathematical Education
• /
• v.54 no.2
• /
• pp.143-165
• /
• 2015

The purpose of this study is to find out the level relevance of the kindergarten curriculum in terms of the kindergarten and elementary school curriculum articulation. For this purpose, a model was developed to assess the level relevance of the curriculum. Next, the achievement standards of the curriculum were analyzed by using this model. Finally, teachers' guidebooks were analyzed, too. The following results were obtained from the analysis. First, five of the 14 achievement standards are rated as 'relevant', and nine of them were 'irrelevant'. Also, six of the irrelevant achievement standards were rated as 'overlap', two of them were rated as 'retrogression', and one of them was rated as 'gap'. I found a lot of problems with the level relevance in the kindergarten curriculum. As the results to analyze teachers' guidebooks, I found that there were the great frequency difference in the activities of teachers' guidebooks.

• Journal of the Korean Society for Library and Information Science
• /
• v.50 no.3
• /
• pp.361-381
• /
• 2016

This study proposed two topical relevance models, simple and complex models, using EEG/ERP techniques. In the simple model regarding simple search tasks, N300 and P3b components are used. The N300 is specific to the semantic processing of pictures and the P3b reflects mechanisms involved in the decision about whether an external stimulus matches or does not match an internal representation of a specific category. In the complex model regarding complex search tasks, on the other hand, N400 and P600 components are used. The N400 reflects activation of an amodel system that integrates both image-based and conceptual representations into a context, whereas the P600 is related to complex cognitive processes. Our research results can be used as a source to design an EEG-based interactive multimedia system.

• MALSORI
• /
• no.46
• /
• pp.117-126
• /
• 2003

This paper presents our language model adaptation for Korean spontaneous speech recognition. Korean spontaneous speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpus. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf/sup */idf similarity. In addition to relevance weighting, we use disfluencies as Predictor to the neighboring words. The best result reduces 9.7% word error rate relatively and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor also.

• Geomechanics and Engineering
• /
• v.1 no.3
• /
• pp.259-262
• /
• 2009

This paper examines the potential of relevance vector machine (RVM) in prediction of pullout capacity of small ground anchors. RVM is based on a Bayesian formulation of a linear model with an appropriate prior that results in a sparse representation. The results are compared with a widely used artificial neural network (ANN) model. Overall, the RVM showed good performance and is proven to be better than ANN model. It also estimates the prediction variance. The plausibility of RVM technique is shown by its superior performance in forecasting pullout capacity of small ground anchors providing exogenous knowledge.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.83-86
• /
• 2003

This paper presents our style-based language model adaptation for Korean conversational speech recognition. Korean conversational speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpora. For style-based language model adaptation, we report two approaches. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf*idf similarity. In addition to relevance weighting, we use disfluencies as predictor to the neighboring words. The best result reduces 6.5％ word error rate absolutely and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor.

• Korean Journal of Cognitive Science
• /
• v.3 no.1
• /
• pp.113-137
• /
• 1991

Coherence and relevaned concern the ways in which discouse is interpreted.In this inversigation, the notions of coherence and relevance are discussed with examples in which they account for discourse interpretation and Charolles's(1983)procedural model of coherence is compared and combined with Sperber & Wilson's(1986a)Principle of Relevance. It is shown that the notion of coherence is not adeqiately established, but that the procedural approach will be a bridge between the prinxiples of coherence and relevance, each of which still remainas to be enriched and completed.

• Health Policy and Management
• /
• v.11 no.2
• /
• pp.123-140
• /
• 2001

Hypotheses In health services research are becoming increasingly more complex and specific. As a result, health services research studies often include multiple independent, intervening, and dependent variables in a single hypothesis. Nevertheless, the statistical models adopted by health services researchers have failed to keep pace with the increasing complexity and specificity of hypotheses and research designs. This article introduces a statistical model well suited for complex and specific hypotheses tests in health services research studies. The covariance structure modeling(CSM) methodology is especially applied to regional relevance indices(RIs) to assess the impact of health resources and healthcare utilization. Data on secondary statistics and health insurance claims were collected by each catchment area. The model for RI was justified by direct and indirect effects of three latent variables measured by seven observed variables, using ten structural equations. The resulting structural model revealed significant direct effects of the structure of health resources but indirect effects of the quantity on RIs, and explained 82% of correlation matrix of measurement variables. Two variables, the number of beds and the portion of specialists among medical doctors, became to have significant effects on RIs by being analyzed using the CSM methodology, while they were insignificant in the regression model. Recommendations for the CSM methodology on health service research data are provided.