• Journal of Intelligence and Information Systems
• /
• v.17 no.2
• /
• pp.109-117
• /
• 2011

The exhibition industry, as technology-intensive, eco-friendly industry, contributes to regional and national development and enhancement of its image as well, if it joins cultural and tourist industry. Therefore, We need to revitalize the exhibition industry, as actively holding an exhibition event. However, to attract a number of exhibition audience, the work of enhancing audience satisfaction and awareness of value for participation should be prioritized after improving quality of service within exhibition hall. As one way to enhance the quality of service, it is thought that the way providing personalized service geared toward each audience is needed. that is, if audience avoids the complexity in exhibition space and it affords them service to enable effective time and space management, it will improve the satisfaction. All such personalized service affordable lets the audience's preference on the basis of each audience profile registered in advance online grasp. and Based on this information, it is provided with exhibition-related information suited their purpose that is the booth for the interesting audience, the shortest path to go to the booth and event via audience's smart phone. and it collects audience's reaction information, such as visiting the booth, participating the event through offered the information in this way and location information for the flow of movement, the present position so that it makes revision of existing each audience profile. After correcting the information, it extracts the individual's preference. hereunder, it provides recommend booth and event information. in other words, it provides optimal information for individual by amendment based on reaction information about recommending information built on basic profile. It provides personalized service dynamic and interactive with audience. This paper will be able to provide the most suitable information for each audience through circular and interactive structure and designed smart-phone application supportable for updating dynamic and interactive personalized service that is able to afford surrounding information in real time, as locating movement position through sensing. The proposed application collects user‘s context information and carrys information gathering function collecting the reaction about searched or provided information via sensing. and it also carrys information gathering function providing needed data for user in exhibition hall. In other words, it offers information about recommend booth of position foundation for user, location-based services of recommend booth and involves service providing detailed information for inside exhibition by using service of augmented reality, the map of whole exhibition as well. and it is also provided with SNS service that is able to keep information exchange besides intimacy. To provide this service, application is consisted of several module. first of all, it includes UNS identity module for sensing, and contain sensor information gathering module handling and collecting the perceived information through this module. Sensor information gathered like this transmits the information gathering server. and there is exhibition information interfacing with user and this module transmits to interesting information collection module through user's reaction besides interface. Interesting information collection module transmits collected information and If valid information out of the information gathering server that brings together sensing information and interesting information is sent to recommend server, the recommend server makes recommend information through inference with gathered valid information. If this server transmit by exhibition information process, exhibition information process module is provided with user by interface. Through this system it raises the dynamic, intelligent personalized service for user.

• Journal of the Daesoon Academy of Sciences
• /
• v.42
• /
• pp.201-226
• /
• 2022

Character education begins with a clear definition of character. There are claims that humanity is reducible to human nature or personality, but when approached from an educational perspective, human nature soon boils down to the question of its ultimate goodness or wickedness. There is a significant difference between the educational systems that emerge from the standpoint that human nature is evil versus the standpoint that human nature is good. With regards to educational outcomes, this can be observed both in terms of human nature and personality. Modern education today sees education as leading the immature to a mature state from the standpoint that human nature is evil. But if human nature is evil, how could we implement an education that would render it good? If character education becomes a system of etiquette education or one of instilling manners whereby simple wrong cases are righted, it would be nothing but a follow-up to the wrong educational cases that had been carried out previously. In that sense, character education is correction; not education. Education should be done in a way that realizes and understands the perfect self rather than unfolding as a process of constantly correcting and reinforcing immature human beings. In that sense, this paper posits that enabling students to understand their own emotions would serve as a correct form of character education. This would be a system of focusing on emotions that reveal the goodness of human nature. Personality can be educated, but education at this time should be a way to bring out a student's already good and even perfected nature. This is more realistic than replacing a 'faulty' character with a good character which supposedly did not exist previously. If personality education morphs into 'emotional self-understanding,' contemplations on 'why not to do' unsavory acts rather than mere negative commands 'don't do that,' and listening to what one's emotions intuit prior to and after given actions, then that would arise to the true standard of a good education.

• Journal of the Korean Society for Library and Information Science
• /
• v.24
• /
• pp.289-332
• /
• 1993

In china, six bibliographies of offical historical books are evaluated at the most important things among the systematically-editing bibliographies. These bibliographies would be usful to study the orign of classical sciences and their development, bibliographic research of Chinese classics, bibliographic judgement on genuine books, titles, authors, volumes. They could be refered to research into graving, correcting, and existence of ancient books. therefore, these bibliographies would be applied to estimation the phase of scientific and cultural development. The study of these bibliographies has been not yet made in Korea. This thesis lays its importance on the background of their appearance, their classification norms, organizing system of their catalogue, and comparison between their difference. 1. Editing and compiling of Chilyak (칠약) by Liu Chin (유흠) and official histories played an important role of entering an apperance of historical book's bibliographies. Chilyak has been lost. However, its classification and compiling system of classical books would be traced by Hansoyemunji(한서예문지) of which basic system is similar to Chilyak. It classified books according to their scientific characteristic. If a few books didn't have their own categories, they were combined by the circles parallel to the books' characteristic. With the books classified under the same scientific characteristic, they were again divided into the scientific schools or structures. It also arranged the same kinds of books according to the chronology. The some books wi th duplicate subjects were classified multiplely by their duplicate subject. 2. Ssu-ma Chon's (사마천) The Historical Records (Saki, 사기) and Pan Ku's (반고) The History of the Former Han Dynasty (Hanso, 한서) has also took effects on appearance of historical books' bibliographies. Covering overall history, Saki was structured by the five parts: The basic annals(본기), the chronological tables (표), the documents (서), the hereditary houses (세가), biographies (열전). The basic annals dealt with kings and courts' affairs according to the chronology. The chronological tables was the records of the annals. The documents described overall the social and cultural systems. The hereditary houses recorded courts' meritorious officials and public figures. The biographies showed exemplars of seventy peoples selected by their social status. Pan Ku(반구)'s The History of the Former Han Dynasty(한서) deserved to be called the prototype for the offical histories after Saki's (사기; The Historical Records) apperance. Although it modelled on Saki, it had set up its own cataloguing system. It was organized by four parts; the basic annals (본기), the chronological tables (표), treatises(지), biographies (열전). The documents in the Hanso(한서) was converted into treatises(지). The hereditary houses and biographies were merged. For the first time, the treatise with The Yemunji could operate function for historical bibliographies. 3. There were six historical bibliographies: Hansoyemunji(한서예문지), Susokyongjeokji (수서경적지), Kudangsokyongjeokji(구당서경적지), Shindangsoyemunji (신당서예문지), Songsayemunji (송사예문지), Myongsayemunji (명사예문지). 1) Modelling on Liu Chin's Chilyak except Chipryak(집략), Hansoyemunji divided the characteristic of the books and documents into six parts: Yukrye(육예), Cheja(제자), Shibu(시부), Pyongsoh(병서), Susul(수술), Pangki(방기). Under six parts, there were thirty eight orders in Hansoyemunji. To its own classification, Hansoyemunji applied the Chilyak's theory of classification that the books or documents were managed according to characteristic of sciences, the difference of schools, the organization of sentences. However the overlapped subjects were deleted and unified into one. The books included into an unsuitable subject were corrected and converted into another. The Hansoyemunji consisted of main preface (Taesoh 대서), minor preface (Sosoh 소서) , the general preface (Chongso 총서). It also recorded the introduction of books and documents, the origin of sciences, the outline of subjects, and the establishment of orders. The books classified by the subject had title, author, and volumes. They were rearranged by titles and the chronological publication year. Sometimes author was the first access point to catalogue the books. If it was necessary for the books to take footnotes, detail notes were formed. The Volume number written consecutively to order and subject could clarify the quantity of books. 2) Refering to Classfication System by Seven Norms (칠분법) and Classification System by Four Norms(사분법), Susokyongjeokji(수서경적지) had accomplished the classification by four norms. In fact, its classification largely imitated Wanhyosoh(완효서)'s Chilrok(칠록), Susokyongjeokji's system of classification consisted of four parts-Kyung(경), Sa(사), Cha(자), Chip(칩). The four parts were divided into 40 orders. Its appendix was again divided into two parts, Buddihism and Taiosm. Under the two parts there were fifteen orders. Totally Susokyongjeokji was made of six parts and fifty five orders. In comparison with Hansoyemunji(한서예문지), it clearly showed the conception of Kyung, Sa, Cha, Chip. Especially it deserved to be paid attention that Hansoyemunji laied history off Chunchu(춘추) and removed history to Sabu(사부). However Chabu(사부) put many contrary subjects such as Cheja(제자), Kiye(기예), Sulsu(술수), Sosol(소설) into the same boundary, which committed errors insufficient theoretical basis. Anothor demerit of Susokyongjeokji was that it dealt with Taiosm scriptures and Buddism scriptures at the appendix because they were considered as quasi-religion. Its compilation of bibliographical facts consisted of main preface(Taesoh 대서), minor preface(Sosoh 소서), general preface (Chongsoh 총서), postscript (Husoh 후서). Its bibliological facts mainly focused on the titles. Its recorded authors' birth date and their position. It wrote the lost and existence of books consecutive to total number of books, which revealed total of the lost books in Su Dynasty. 3) Modelling on the basis of Kokumsorok(고분서록) and Naewaekyongrok(내외경록), Kudangsokyongjeokji(구당서경적지) had four parts and fourty five orders. It was estimated as the important role of establishing basic frame of classification by four norms in classification theory's history. However it had also its own limit. Editing and compling orders of Kudangsokyongjeokji had been not progressively changed. Its orders imitated by and large Susokyongjeokji. In Its system of organizing catalogue, with its minor preface and general preface deleting, Kudangsokyongjeokji by titles after orders sometimes broke out confusion because of unclear boundaries between orders. 4) Shindangsoyemunji(신당서예문지), adding 28,469 books to Kudangsokyongjeokji, recorded 82,384 books which were divided by four parts and fourty four orders. In comparison with Kudangkyongjeokj, Sindangsoyemunji corrected unclear order's norm. It merged the analogical norms four orders (for instance, Kohun 고훈 and Sohakryu 소학류) and seperated the different norms four orders (for example, Hyokyong 효경 and Noneuhryu 논어류, Chamwi 참위 and Kyonghaeryu 경해류, Pyonryon 편년 and Wisaryu 위사류). Recording kings' behaviors and speeches (Kikochuryu 기거주류) in the historical parts induced the concept of specfication category. For the first time, part of Chipbu (집부) set up the order of classification norm for historical and literatural books and documents (Munsaryu 문사류). Its editing and compiling had been more simplified than Kudangsokyongjeokji. Introduction was written at first part of bibliographies. Appendants except bibliographic items such subject, author, title, volume number, total were omitted. 5) Songsayemunji(송사예문지) were edited in the basis of combining Puksong(북송) and Namsong(남송), depending on Sabukuksayemunji(사부국사예문지). Generally Songsayemunji had lost a lot of bibliographical facts of many books. They were duplicated and wrongly classified books because it committed an error of the incorrectly annalistic editing. Particularly Namsong showed more open these defaults. Songsayemunji didin't include the books published since the king Youngchong(영종). Its system of classification was more better controlled. Chamwiryu(참위류) in the part of Kyongbu(경부) was omitted. In the part of history(Sabu 사부), recordings of kings' behaviors and speeches more merged in the annals. Historical abstract documents (Sachoryu 사초류) were seperately arranged. In the part of Chabu(자부), Myongdangkyongmaekryu(명당경맥류) and Euisulryu(의술류) were combined. Ohangryu(오행류) were laied off Shikuryu(시구류). In the part of Chipbu(집부), historical and literatural books (Munsaryu 문사류) were independentely arranged. There were the renamed orders; from Wisa(위사) to Paesa(패사), Chapsa (잡사) to Pyolsa(열사), Chapchonki(잡전기) to Chonki(전기), Ryusoh(류서) to Ryusa(류서). Introduction had only main preface. The books of each subject catalogued by title, the volume number, and author and arranged mainly by authors. Annotations were written consecutively after title and the volume number. In the afternote the number of not-treated books were revealed. Difference from Singdangsohyemunji(신당서예문지) were that the concept and boundary of orders became more clearer. It also wrote the number of books consecutive to main subject. 6) Modelling on Chonkyongdangsomok (경당서목), Myongsayemunji(명사예문지) was compiled in the basis of books and documents published in the Ming Danasty. In classification system, Myongsayemunji partly merged and the seperated some orders for it. It also deleted and renamed some of orders. In case of necessity, combining of orders' norm was occured particulary in the part of Sabu(사부) and Chabu(자부). Therefore these merging of orders norm didn't offer sufficient theretical background. For example, such demerits were seen in the case that historical books edited by annals were combined with offical historical ones which were differently compiled and edited from the former. In the part of Chabu(자부), it broke out another confusion that Pubga(법가), Meongga(명가), Mukga(묵가), Chonghweongka's(종횡가) thoughts were classified in the Chapka(잡가). Scriptures of Taiosim and Buddhism were seperated from each other. There were some deleted books such as Mokrokryu(목록류), Paesaryu(패사류) in the part of history (Sabu 사부) and Chosaryu(초사류) in the part of Chipbu(집부). The some in the each orders had been renamed. Imitating compiling system of Songsayemunji(송사예문지), with reffering to its differ-ence, Myongsayemunji(명사예문지) wrote the review and the change of the books by author. The number of not-treated books didn't appear at the total. It also deleted the total following main subject.

• Journal of the Korean Society of Radiology
• /
• v.7 no.1
• /
• pp.17-24
• /
• 2013

This study estimate radiation biological danger factor by measuring patient's exposed dose and propose the low way of patient's exposed dose in panoramic radiography. We seek correcting constant of OSL dosimeter for minimize the error of exposed dose's measurement and measure the Left, Right crystalline lens, thyroid, directly included upper, lower lips, the maxillary bone and the center of photographing that indirect included in panoramic radiography by using the human body model standard phantom advised in ICRP. In result, the center of photographing's level of radiation maximum value is $413.67{\pm}6.53{\mu}Gy$ and each upper, lower lips is $217.80{\pm}2.98{\mu}Gy$, $215.33{\pm}2.61{\mu}Gy$. Also in panoramic radiography, indirect included Left, Right crystalline lens's level of radiation are $30.73{\pm}2.34{\mu}Gy$, $31.87{\pm}2.50{\mu}Gy$, and thyroid's level of measured exposed dose can cause effect of radiation biological and we need justifiable analysis about radiation defense rule and substantiation advised international organization for the low way of patient's exposed dose in panoramic radiography of dental clinic and we judge need the additional study about radiation defense organization for protect the systematize protocol's finance and around internal organs for minimize until accepted by many people that is technological, economical and social fact by using panoramic measurement.

• Information Systems Review
• /
• v.7 no.2
• /
• pp.229-255
• /
• 2005

The Korea Development Bank promoted a total outsourcing for IT operation in 1999 for the first time in the banking industry. The Korea Development Bank became the center of public attention because the most banks were unwilling to take an outsourcing with external sources for the reason of financial operation accidents, securities, and threats of strikes. After the introduction of the total IT outsourcing, the Korea Development Bank has continuously diagnosed the problems of the IT outsourcing and adopted various proper complements for the enhancement of the IT outsourcing. As the result of the enhancement, the IT outsourcing of the Korea Development Bank marched into the joint liability operation period after going through the outsourcing operation period and the co-operation period. The joint liability operation which is the most leading outsourcing system which is adopted by the Korea Development Bank for the first time in the banking industry. Through the joint liability operation, the Korea Development Bank could accept the most up-to-date IT, concentrate internal manpower on the core capability, and secure flexibility of manpower. Also, the bank changed the relationship between the bank and the external sources from the one-sided relationship between a producers and a consumer to the joint liability relationship on which both sides are responsible for the operation, and could integrate the internal capacity with the professional know-how of the external IT outsourcing company. In this paper, we testified the soundness and validity for the worries of banks about the total IT outsourcing with external sources. And, we arranged the advantages and outcomes of the total IT outsourcing with external sources compared to the IT outsourcing with internal sources. Moreover, we expect that we can improve the closed financial IT outsourcing industry structure and raise the world competitive power of domestic IT outsourcing companies by correcting wrong ideas on the IT outsourcing with external sources.

• Journal of the Korean earth science society
• /
• v.36 no.1
• /
• pp.109-124
• /
• 2015

A novel approach, hybrid surface rainfall (KNU-HSR) technique developed by Kyungpook Natinal University, was utilized for improving the radar rainfall estimation. The KNU-HSR technique estimates radar rainfall at a 2D hybrid surface consistings of the lowest radar bins that is immune to ground clutter contaminations and significant beam blockage. Two HSR techniques, static and dynamic HSRs, were compared and evaluated in this study. Static HSR technique utilizes beam blockage map and ground clutter map to yield the hybrid surface whereas dynamic HSR technique additionally applies quality index map that are derived from the fuzzy logic algorithm for a quality control in real time. The performances of two HSRs were evaluated by correlation coefficient (CORR), total ratio (RATIO), mean bias (BIAS), normalized standard deviation (NSD), and mean relative error (MRE) for ten rain cases. Dynamic HSR (CORR=0.88, BIAS= $-0.24mm\;hr^{-1}$, NSD=0.41, MRE=37.6%) shows better performances than static HSR without correction of reflectivity calibration bias (CORR=0.87, BIAS= $-2.94mm\;hr^{-1}$, NSD=0.76, MRE=58.4%) for all skill scores. Dynamic HSR technique overestimates surface rainfall at near range whereas it underestimates rainfall at far ranges due to the effects of beam broadening and increasing the radar beam height. In terms of NSD and MRE, dynamic HSR shows the best results regardless of the distance from radar. Static HSR significantly overestimates a surface rainfall at weaker rainfall intensity. However, RATIO of dynamic HSR remains almost 1.0 for all ranges of rainfall intensity. After correcting system bias of reflectivity, NSD and MRE of dynamic HSR are improved by about 20 and 15%, respectively.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.3 no.2
• /
• pp.141-151
• /
• 1993

Blood samples obtained from 200 adults who had visited the "S" general hospital were analyzed to compare the zinc protoporphyrin (ZPP) levels quantified by high performance liquid chromatograph (HPLC) and by hematofluorometer (HF) to investigate the methodological difference if any and the relationship between the levels of blood lead and ZPP among no-lead exposed adults. Also investigated were the distribution of ZPP and protoporphyrin IX (PPIX) concentrations, the establishment of normal levels of blood ZPP and blood lead, and the contribution of age and sex factors to these values. These subjects had no previous occupational exposure to lead. The results obtained were as follows : 1. The mean values of blood lead for male and female subjects were $9.46{\pm}2.44{\mu}g/dl$ and $8.09{\pm}2.17{\mu}g/dl$, respectively. The difference observed in the mean concentrations between male and female subjects was statistically very significant. 2. The mean values of blood ZPP by HPLC for male and female subjects were $15.94{\pm}4.55{\mu}g/dl$ and $22.26{\pm}6.61{\mu}g/dl$, respectively. The difference observed in the mean concentrations between male and female subjects was statistically not significant. The mean values of blood PPIX by HPLC for male and female subjects were $2.51{\pm}1.78{\mu}g/dl$ and $2.81{\pm}1.56{\mu}g/dl$, respectively. The difference observed in the mean concentrations between male and female subjects was statistically not significant. 3. The mean values of blood ZPP by HF for male and female subjects were $28.44{\pm}7.11{\mu}g/dl$ and $37.77{\pm}8.04{\mu}g/dl$, respectively. The difference observed in the mean concentrations between male and female subjects was statistically very significant. 4. No statistically significant correlation was found between the levels of blood ZPP and blood lead. 5. The ratio of ZPP and protoporphyrin IX (PPIX) concentration to erythrocyte protoporphyrin (EP, EP=ZPP+PPIX) concentration was 87.4% and 12.6%, respectively. 6. A statistically very significant correlation was found between the ZPP concentrations determined by HPLC and the values by HF (r=0.7565). The ZPP concentraitons quantified by HF were 1.75 times as high as the values obtained by HPLC. 7. The blood ZPP concentrations quantified by HPLC, HF, and spectrofluorometer (SF) from the blood samples obtained from 14 lead-exposed workers and from 16 no-lead exposed adults showed wide variations. The ZPP concentrations by HF were the highest followed by the levels obtained by SF and by HPLC. In the exposed group, no statistically significant difference was found among three methods of quantifying blood ZPP levels. In the no-lead exposed group, however, statistically significant difference was observed among these methods. The ZPP concentrations by HF were about twice as high as those of by HPLC or by SF. Among three methods of quantifying blood ZPP (HPLC, SF and HF), the results revealed significant difference. Therefore it is suggested that objective methods of quantifying blood ZPP and a system of correcting different ZPP levels be developed by the ministry of Labor.

• Asia pacific journal of information systems
• /
• v.19 no.2
• /
• pp.157-178
• /
• 2009

Corporate credit rating is a very important factor in the market for corporate debt. Information concerning corporate operations is often disseminated to market participants through the changes in credit ratings that are published by professional rating agencies, such as Standard and Poor's (S&P) and Moody's Investor Service. Since these agencies generally require a large fee for the service, and the periodically provided ratings sometimes do not reflect the default risk of the company at the time, it may be advantageous for bond-market participants to be able to classify credit ratings before the agencies actually publish them. As a result, it is very important for companies (especially, financial companies) to develop a proper model of credit rating. From a technical perspective, the credit rating constitutes a typical, multiclass, classification problem because rating agencies generally have ten or more categories of ratings. For example, S&P's ratings range from AAA for the highest-quality bonds to D for the lowest-quality bonds. The professional rating agencies emphasize the importance of analysts' subjective judgments in the determination of credit ratings. However, in practice, a mathematical model that uses the financial variables of companies plays an important role in determining credit ratings, since it is convenient to apply and cost efficient. These financial variables include the ratios that represent a company's leverage status, liquidity status, and profitability status. Several statistical and artificial intelligence (AI) techniques have been applied as tools for predicting credit ratings. Among them, artificial neural networks are most prevalent in the area of finance because of their broad applicability to many business problems and their preeminent ability to adapt. However, artificial neural networks also have many defects, including the difficulty in determining the values of the control parameters and the number of processing elements in the layer as well as the risk of over-fitting. Of late, because of their robustness and high accuracy, support vector machines (SVMs) have become popular as a solution for problems with generating accurate prediction. An SVM's solution may be globally optimal because SVMs seek to minimize structural risk. On the other hand, artificial neural network models may tend to find locally optimal solutions because they seek to minimize empirical risk. In addition, no parameters need to be tuned in SVMs, barring the upper bound for non-separable cases in linear SVMs. Since SVMs were originally devised for binary classification, however they are not intrinsically geared for multiclass classifications as in credit ratings. Thus, researchers have tried to extend the original SVM to multiclass classification. Hitherto, a variety of techniques to extend standard SVMs to multiclass SVMs (MSVMs) has been proposed in the literature Only a few types of MSVM are, however, tested using prior studies that apply MSVMs to credit ratings studies. In this study, we examined six different techniques of MSVMs: (1) One-Against-One, (2) One-Against-AIL (3) DAGSVM, (4) ECOC, (5) Method of Weston and Watkins, and (6) Method of Crammer and Singer. In addition, we examined the prediction accuracy of some modified version of conventional MSVM techniques. To find the most appropriate technique of MSVMs for corporate bond rating, we applied all the techniques of MSVMs to a real-world case of credit rating in Korea. The best application is in corporate bond rating, which is the most frequently studied area of credit rating for specific debt issues or other financial obligations. For our study the research data were collected from National Information and Credit Evaluation, Inc., a major bond-rating company in Korea. The data set is comprised of the bond-ratings for the year 2002 and various financial variables for 1,295 companies from the manufacturing industry in Korea. We compared the results of these techniques with one another, and with those of traditional methods for credit ratings, such as multiple discriminant analysis (MDA), multinomial logistic regression (MLOGIT), and artificial neural networks (ANNs). As a result, we found that DAGSVM with an ordered list was the best approach for the prediction of bond rating. In addition, we found that the modified version of ECOC approach can yield higher prediction accuracy for the cases showing clear patterns.

• The Journal of Korean Society for Radiation Therapy
• /
• v.35
• /
• pp.41-51
• /
• 2023

Purpose: This study compares and analyzes the image quality of 3D-CBCT(Cone Beam Computed-Tomography) and Gated CBCT according to baseline changes during SBRT(Stereotactic Body RadioTherapy) in lung cancer patients to find a useful CBCT method for correcting movement due to breathing Materials and methods : Insert a solid tumor material with a diameter of 3 cm into the QUASAR^TM phantom. 4-Dimentional Computed-Tomography(4DCT) images were taken with a speed of the phantom at period 3 sec and a maximum amplitude of 20 mm. Using the contouring menu of the computerized treatment planning system Eclipse^TM Gross Tumor Volume was outlined on solid tumor material. Set-up the same as when acquiring a 4DCT image using Truebeam STx^TM, breathing patterns with baseline changes of 1 mm, 3 mm, and 5 mm were input into the phantom to obtain 3D-CBCT (Spotlight, Full) and Gated-CBCT (Spotlight, Full) images five times repeatedly. The acquired images were compared with the Signal-to-Noise Ratio(SNR), Contrast-to-Noise Ratio(CNR), Tumor Volume Length, and Motion Blurring Ratio(MBR) based on the 4DCT image. Results: The average Signal-to-Noise Ratio, Contrast-to-Noise Ratio, Tumor Volume Length and Motion Blurring Ratio of Spotlight Gated CBCT images were 13.30±0.10%, 7.78±0.16%, 3.55±0.17%, 1.18±0.06%. As a result, Spotlight Gated-CBCT images according to baseline change showed better values than Spotligtht 3D-CBCT images. Also, the average Signal-to-Noise Ratio, Contrast-to-Noise Ratio, Tumor Volume Length and Motion Blurring Ratio of Full Gated CBCT images were 12.80±0.11%, 7.60±0.11%, 3.54±0.16%, 1.18±0.05%. As a result Full GatedCBCT images according to baseline change showed better values than Full 3D-CBCT images. Conclusion : Compared to 3D-CBCT images, Gated-CBCT images had better image quality according to the baseline change, and the effect of Motion Blurring Artifacts caused by breathing was small. Therefore, it is considered useful to image guided using Gated-CBCT when a baseline change occurs due to difficulty in regular breathing during SBRT that exposes high doses in a short period of time

• Journal of Intelligence and Information Systems
• /
• v.24 no.2
• /
• pp.59-83
• /
• 2018

With the increasing importance of sentiment analysis to grasp the needs of customers and the public, various types of deep learning models have been actively applied to English texts. In the sentiment analysis of English texts by deep learning, natural language sentences included in training and test datasets are usually converted into sequences of word vectors before being entered into the deep learning models. In this case, word vectors generally refer to vector representations of words obtained through splitting a sentence by space characters. There are several ways to derive word vectors, one of which is Word2Vec used for producing the 300 dimensional Google word vectors from about 100 billion words of Google News data. They have been widely used in the studies of sentiment analysis of reviews from various fields such as restaurants, movies, laptops, cameras, etc. Unlike English, morpheme plays an essential role in sentiment analysis and sentence structure analysis in Korean, which is a typical agglutinative language with developed postpositions and endings. A morpheme can be defined as the smallest meaningful unit of a language, and a word consists of one or more morphemes. For example, for a word '예쁘고', the morphemes are '예쁘(= adjective)' and '고(=connective ending)'. Reflecting the significance of Korean morphemes, it seems reasonable to adopt the morphemes as a basic unit in Korean sentiment analysis. Therefore, in this study, we use 'morpheme vector' as an input to a deep learning model rather than 'word vector' which is mainly used in English text. The morpheme vector refers to a vector representation for the morpheme and can be derived by applying an existent word vector derivation mechanism to the sentences divided into constituent morphemes. By the way, here come some questions as follows. What is the desirable range of POS(Part-Of-Speech) tags when deriving morpheme vectors for improving the classification accuracy of a deep learning model? Is it proper to apply a typical word vector model which primarily relies on the form of words to Korean with a high homonym ratio? Will the text preprocessing such as correcting spelling or spacing errors affect the classification accuracy, especially when drawing morpheme vectors from Korean product reviews with a lot of grammatical mistakes and variations? We seek to find empirical answers to these fundamental issues, which may be encountered first when applying various deep learning models to Korean texts. As a starting point, we summarized these issues as three central research questions as follows. First, which is better effective, to use morpheme vectors from grammatically correct texts of other domain than the analysis target, or to use morpheme vectors from considerably ungrammatical texts of the same domain, as the initial input of a deep learning model? Second, what is an appropriate morpheme vector derivation method for Korean regarding the range of POS tags, homonym, text preprocessing, minimum frequency? Third, can we get a satisfactory level of classification accuracy when applying deep learning to Korean sentiment analysis? As an approach to these research questions, we generate various types of morpheme vectors reflecting the research questions and then compare the classification accuracy through a non-static CNN(Convolutional Neural Network) model taking in the morpheme vectors. As for training and test datasets, Naver Shopping's 17,260 cosmetics product reviews are used. To derive morpheme vectors, we use data from the same domain as the target one and data from other domain; Naver shopping's about 2 million cosmetics product reviews and 520,000 Naver News data arguably corresponding to Google's News data. The six primary sets of morpheme vectors constructed in this study differ in terms of the following three criteria. First, they come from two types of data source; Naver news of high grammatical correctness and Naver shopping's cosmetics product reviews of low grammatical correctness. Second, they are distinguished in the degree of data preprocessing, namely, only splitting sentences or up to additional spelling and spacing corrections after sentence separation. Third, they vary concerning the form of input fed into a word vector model; whether the morphemes themselves are entered into a word vector model or with their POS tags attached. The morpheme vectors further vary depending on the consideration range of POS tags, the minimum frequency of morphemes included, and the random initialization range. All morpheme vectors are derived through CBOW(Continuous Bag-Of-Words) model with the context window 5 and the vector dimension 300. It seems that utilizing the same domain text even with a lower degree of grammatical correctness, performing spelling and spacing corrections as well as sentence splitting, and incorporating morphemes of any POS tags including incomprehensible category lead to the better classification accuracy. The POS tag attachment, which is devised for the high proportion of homonyms in Korean, and the minimum frequency standard for the morpheme to be included seem not to have any definite influence on the classification accuracy.