• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.4
• /
• pp.141-148
• /
• 2013

Automated classification systems which utilize machine learning develops classification models through learning process, and then classify unknown data into predefined set of categories according to the model. The performance of machine learning-based classification systems relies greatly upon the quality of features composing classification models. For textual data, we can use their word terms and structure information in order to generate the set of features. Particularly, in order to extract feature from Web documents, we need to analyze tag and hyperlink information. Recent studies on Web document classification focus on feature engineering technology other than machine learning algorithms themselves. Thus this paper proposes a novel method of incorporating feature selection and weighting which can improves classification models effectively. Through extensive experiments using Web-KB document collections, the proposed method outperforms conventional ones.

• Journal of the Korea Computer Graphics Society
• /
• v.13 no.1
• /
• pp.15-24
• /
• 2007

This study is focused on fellowing two issues for simulating colored paper mosaic. The first one is paper tile generation, and the other one is arrangement. At first we define the paper data as two layered polygon. Then, generate paper tile from paper data, and attach it on the best location of canvas according to energy value. The proposed method in this study has some benefits like follow. First, we can obtain more natural shaped paper tile because the phase is similar to real process of fearing paper. Second, if we redefine the structure of paper like newspaper, magazine and so on, the result can be generated with the variable type of paper. And lastly, by adjusting parameters of energy function, we can control the position of the torn paper.

• Asia Pacific Journal of Business Review
• /
• v.1 no.2
• /
• pp.37-62
• /
• 2017

To analyze the effects of R&D expenditure on the firm value of Korean firms, we classified portfolios based on R&D activity levels. After that, we conducted a time-series analysis to assess excess returns from the portfolios. To carry out such an analysis, an empirical analysis of excess returns in the capital market was performed by using the monthly earning rate of stocks from 2000 to 2013. The purpose of this research is to provide basic data on investment to stakeholders in the capital market by analyzing the effects of R&D on the firm value and to overcome scholarly limitations by offering a new model of analysis. The criteria for classifying the portfolios were based on R&D expenditure levels. The analysis models follow the Fama-French Three-Factor Model and the Carhart Four-Factor Model. The analyses results are as follows. Extrapolating monthly profit rates based on R&D expenditure levels, portfolios with low R&D expenditures showed higher earning rates than those with high R&D expenditures. This suggests that high R&D expenditures did not translate into high earning rates. The investor depreciates the R&D expenditures related profitability and the possibility of success in the market, leading to falls in stock prices and a failure to give a positive effect on the firm value. Our research differs from the previous investigations as we carried out an empirical analysis based on the actual investors' attitudes about R&D expenditures and how these can generate excess earnings. Our research results show that the data related to R&D expenditure are not reflected fully in the market.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.277-280
• /
• 2016

Real-time remote monitoring system has an important value in many surveillance situations. It allows someone to be informed of what is happening in his monitoring locations. Kinect v2 is a new kind of camera which gives computers eyes and can generate different data such as color and depth images, audio input and skeletal data. In this paper, using Kinect v2 sensor with its depth image, we present a monitoring system in a space covered by Kinect. Therefore, based on space covered by Kinect camera, we define a target area to monitor using depth range by setting minimum and maximum distances. With computer vision library (Emgu CV), if there is an object tracked in the target space, kinect camera captures the whole image color and sends it in database and user gets at the same time a notification on his mobile device wherever he is with internet access.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.369-372
• /
• 2018

This paper suggests non-parallel-voice-conversion network conversing voice between unmapped voice pair as source voice and target voice. Conventional voice conversion researches used learning methods that minimize spectrogram's distance error. Not only these researches have some problem that is lost spectrogram resolution by methods averaging pixels. But also have used parallel data that is hard to collect. This research uses PPGs that is input voice's phonetic data and a GAN learning method to generate more clear voices. To evaluate the suggested method, we conduct MOS test with GMM based Model. We found that the performance is improved compared to the conventional methods.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.5
• /
• pp.886-891
• /
• 2015

Recently, The World Health Organization announced that harms human health because of air pollution that are emerging as threats to human health worldwide. according to the, Seoul, 2011 According to a July 2014 - Public Facilities indoor air quality measurements, were examined to be in violation of indoor air pollution standards in most multi-use facility. Indoor air pollution resulting from this is present in the paper, and cause disease, such as pulmonary disease, asthma, bronchitis and to In this connection, the measurement of indoor air quality by using the environment sensor, analyzing the measured data to generate an actuator signal required for ventilation and improve indoor air quality by implementing a monitoring system with real-time measurement, autonomously managing the air quality in our lives so that it can be.

• The KIPS Transactions:PartD
• /
• v.16D no.6
• /
• pp.845-850
• /
• 2009

Use of XML data has been increased with growth of Web 2.0 environment. XML is recognized its advantages by using based technology of RSS or ATOM for transferring information from blogs and news feed. Bitmap clustering is a method to keep index in main memory based on Relational DBMS, and which performed better than the other XML indexing methods during the evaluation. Existing method generates too many clusters, and it causes deterioration of result of searching quality. This paper proposes k-Bitmap clustering method that can generate user defined k clusters to solve above-mentioned problem. The proposed method also keeps additional inverted index for searching excluded terms from representative bits of k-Bitmap. We performed evaluation and the result shows that the users can control the number of clusters. Also our method has high recall value in single term search, and it guarantees the searching result includes all related documents for its query with keeping two indices.

• Journal of the Korean Society of Costume
• /
• v.65 no.2
• /
• pp.157-175
• /
• 2015

The study was aimed to develop a torso-type dress form representing body features of the female fashion models in Korea. To fulfill this purpose, 5 female fashion models aged between 20 and 26 having the average body measurements of professional fashion models in Korea were selected and their 3-D whole body scanned data were analysed. The 3-D whole body scanning method enabled to generate a virtual female fashion model within the CAD system by measuring the subjects' body shapes and sizes. In addition, the virtual model's body data led the development of a standard female fashion model dress form for the efficient fashion show preparation. In order to manufacture the real dress form for female fashion models, 3-D printing technology was adopted. The consequent results are as follows: (1) the body measurements (unit: cm) of the developed dress form were: biacromion length, 36.0, bust point to bust point, 16.6, front/back interscye lengths, 32.0/33.0, neck point to breast point, 26.0, neck point to breast point to waist line, 41.5, waist front/back lengths, 34.5/38.5, waist to hip length, 24.0, bust circumference, 85.0, underbust circumference, 75.0, waist circumference, 65.0, hip circumference, 92.0. (2) the body measurements differences between the developed and existing dress forms were highlighted with the body measurements of neck point to breast point and waist to hip length. (3) the body shape features of the developed dress form showed that bust, shoulder blade, shoulder slope, abdomen and back waist line to hip line parts were more realistically manufactured.

• Proceedings of KOSOMES biannual meeting
• /
• 2005.05a
• /
• pp.135-140
• /
• 2005

Currently, ECDIS and GPS plotters are much used as equipment providing operators with route information, but they do not have any function of automatic route creation and route explanation, so available for only experienced operators. Especially, the present situation is tint no study is made of the automatic route creation and route explanation using ENC for ECDIS, substitution system of paper chart. ENC is the electronic navigation chart that is produced using S-52, S-57 standard format required by IHQ. In this paper, an Vessel Operator Support System(VOSS) is proposed to generate an optimal route where ENC and GPS data is fusioned induding the wind direction and speed of an anemometer and tide data. The proposed system was testified by a simulation, and its effectiveness was verified.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.38 no.3
• /
• pp.39-48
• /
• 2015

In this study, granting the optimal loan limit on SME (Small and Medium Enterprise) loans of financial institutions was proposed using the traditional newsvendor model. This study was the first domestic case study that applied the newsvendor model that was mainly used to calculate the optimum order quantity under some uncertain demands to the calculation of the loan limit (debt ceiling) of institutions. The method presented in this study made it possible to calculate the loan limit (debt ceiling) to maximize the revenue of a financial institution using probability functions, applied the newsvendor model setting the order volume of merchandise goods as the loan product order volume of the financial institution, and proposed, through the analysis of empirical data, the availability of additional loan to the borrower and the reduction of the debt ceiling and a management method for the recovery of the borrower who could not generate profit. In addition, the profit based loan money management model presented in this study also demonstrated that it also contributed to some extent to the prediction of the bankruptcy of the borrowing SME (Small and Medium Enterprise), as well as the calculation of the loan limit based on profit, by deriving the result values that the borrowing SME (Small and Medium Enterprise) actually went through bankruptcy at later times once the model had generated a signal of loan recovery for them during the validation of empirical data. accordingly, The method presented in this study suggested a methodology to generated a signal of loan recovery to reduce the losses by the bankruptcy.