• Speech Sciences
• /
• v.14 no.2
• /
• pp.127-135
• /
• 2007

Feature vectors of speech are generally extracted from whole frequency domain. The inherent character of a speaker is located in the low band or high band frequency. However, if the speech is corrupted by narrowband noise with concentrated energy, speaker verification performance is reduced as the individual characteristic is removed. In this paper, we propose a PCA Covariance Model based on the multiband to extract the robust feature vectors against the narrowband noise. First, we divide the overall frequency band into several subbands. Second, the correlation of feature vectors extracted independently from each subband is removed by PCA. The distance obtained from each subband has different distribution. To normalize against the different distribution, we moved the value into the normalized distribution through the mapping function. Finally, the represented value applying the weighting function is used for speaker verification. In the experiments, the proposed method shows better performance of the speaker verification and reduces the computation.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.14 no.5
• /
• pp.549-560
• /
• 1989

In this paper, the method of determining the initial weighting vlaues fuctioning as a filter under the Directional Constrained Minimization of Power(DCMP) algorithm is presented. By analyzing the sideband beamformer with the Finite Impulse Response (FIR) filter concepts, the constraints of any desired directions are obtained and the initial weighing values with fast adaptation time are formulated from those constraints. By applying this proposed initial weighting values to the DCMP and the spatial averaging processor, the interference of a desired direction and the coherent noises are eliminated at the same time. The improvement of this method compared with the existing algorithm is confirmed by computer simulation.

• Journal of the Society of Naval Architects of Korea
• /
• v.54 no.5
• /
• pp.368-377
• /
• 2017

Outfittings of offshore plants and high value-added vessels, such as FPSO, drillship, are much more than outfittings of general ships. So the complexity of the outfittings is increased and the importance of outfittings has also increased. But, there is no clear method to define the installation sequence of outfittings, yet. In this study, we developed the algorithm to define the installation sequence of outfittings by applying variables and constraints related to outfitting, such as process parameters, weighting coefficients, installation constraints. Also, we developed the application that applied the algorithm and compared cases by changing the weighting coefficients of process parameters. We verified the practicality of the algorithm by developing the application. The results of this study are that the accuracy of the outfitting planning is improved and efficient lead time can be predicted by defining the installation sequence of outfittings.

• Atmosphere
• /
• v.24 no.2
• /
• pp.235-243
• /
• 2014

Sources of aviation turbulence vary through the seasons, especially in the East Asia including Korean peninsula, associated primarily with the changes in the jet/front system and convective activities. For this reason, a seasonal Korean aviation Turbulence Guidance (KTG) system (seasonal-KTG) is developed in the present study by using pilot reports (PIREPs) and analysis data of the operational Unified Model (UM) of the Korea Meteorological Administration (KMA) for two years between June 2011 and May 2013. Twenty best diagnostics of aviation turbulence in each season are selected by the method of probability of detection (POD) using the PIREPs and UM data. After calculating a weighting value of each selected diagnostics using their area under curve (AUC), the 20 best diagnostics are combined with the weighting scores into a single ensemble-averaged index by season. Compared with the current operational-KTG system that is based on the diagnostics applying all seasons, the performances of the seasonal-KTG system are better in all seasons, except in fall.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.5
• /
• pp.61-69
• /
• 2014

Collaborative filtering recommendation creates similar item subset or similar user subset based on user preference about items and predict user preference to particular item by using them. Thus, if preference matrix has low density, reliability of recommendation will be sharply decreased. To solve these problems we suggest Hybrid Preference Prediction Technique Using Weighting based Data Reliability. Preference prediction is carried out by creating similar item subset and similar user subset and predicting user preference by each subset and merging each predictive value by weighting point applying model condition. According to this technique, we can increase accuracy of user preference prediction and implement recommendation system which can provide highly reliable recommendation when density of preference matrix is low. Efficiency of this system is verified by Mean Absolute Error. Proposed technique shows average 21.7% improvement than Hao Ji's technique when preference matrix sparsity is more than 84% through experiment.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.31 no.1
• /
• pp.49-55
• /
• 2013

In this paper, we proposed a good classifier to match different spatial data sets by applying evaluation of classifiers performance in data mining and biometrics. For this, we calculated distances between a pair of candidate features for matching criteria, and normalized the distances by Min-Max method and Tanh (TH) method. We defined classifiers that shape similarity is derived from fusion of these similarities by CRiteria Importance Through Intercriteria correlation (CRITIC) method, Matcher Weighting method and Simple Sum (SS) method. As results of evaluation of classifiers performance by Precision-Recall (PR) curve and area under the PR curve (AUC-PR), we confirmed that value of AUC-PR in a classifier of TH normalization and SS method is 0.893 and the value is the highest. Therefore, to match different spatial data sets, we thought that it is appropriate to a classifier that distances of matching criteria are normalized by TH method and shape similarity is calculated by SS method.

• Journal of Internet Computing and Services
• /
• v.2 no.5
• /
• pp.9-19
• /
• 2001

In this paper. we develop associate knowledge extraction method for finding and expanding user preference knowledge automatically from web document database. To reflect user interest or preferences, agent explores and extracts relevant information to central term involving the intent of users from the example documents. To do so, we apply association rule exploration data-mining method to the extraction of the relevant objects in the web documents. Also, to give the weighted-value to the extracted and relevant information, we present associate tag block-based weighting method. We applied to bioinformatics above associate knowledge extraction method to find related keywords.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.9 no.2
• /
• pp.23-32
• /
• 2013

The purpose of this study is to make effective mobile grid considered general environment, which can be summarized as irregular mobility, service exploration, data sharing, variety of machines, limit to the battery duration, etc. The data was extracted from the Dartmouth College. We analysed mobile use pattern of a specific group and applied pattern using hybrid method. As a result, we could adjust infra usage effectively and appropriately and cost cutting and increase satisfaction of user. In this study, by applying weighting method based on access time interval, we analysed use pattern added time variation with association rule during users in mobile grid environment. We proposed more stable way to manage patterns in a mobile grid environment that is being used as a hybrid form to process the data value received from the server in real time. Further studies are needed to get appropriate use pattern by group using use patterns of various groups.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.12 no.5
• /
• pp.797-804
• /
• 2017

A social network is a typical social science environment with both network and iteration characteristics. This research presents a reaction analysis model of how each node responds to social networks when given input such as promotions or incentives. In addition, the setting value of a specific node is changed while examining the response of each node. And we try to understand the reactions of the nodes involved. The reaction analysis model is constructed by applying various techniques such as unidirectional, fuzzy set, weighting, and cyclic feedback, so it can accommodate the complicated environment of practice. Finally, the implementation model is implemented using Vensim rather than NetLogo because it requires repetitive input, change of setting value in real time, and analysis of association between nodes.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.05a
• /
• pp.234-240
• /
• 2003

This study aims to investigate dynamic properties of express buses in the very low frequencies which affect motion sickness incidence. Since passengers often use express buses for long distance traveling, it is a critical point whether a give rise to motion sickness or not. In the study accelerations at the three points on the floor of the six test vehicles were measured during the driving at constant speeds. By applying frequency weighting curves suggested in ISO 2631-1 and ISO 2631-3, physical amount of accelerations were changed into perceptual amount which determines incidence of motion sickness. Motion sickness dose values were calculated from the frequency weighted time history of accelerations, and compared between the vehicles, driving conditions, and the seat positions in the bus. During the driving on public road and high ways for 50 minutes vomiting incidence ratios ranged 0.4 to 0.8%, which were equivalent to 2.4 to 4.8% for 5 hours' driving. The value of 4.8 % means two among 45 passengers may vomit after the traveling, which is very serious situation. Considering the very smooth driving condition at which the data were collected, motion sickness dose values will increase in real situations