• Journal of KIISE:Software and Applications
• /
• v.26 no.10
• /
• pp.1219-1229
• /
• 1999

일반적으로 AND, OR, NOT과 같은 연산자를 사용하는 불리언 질의는 사용자의 검색의도를 정확하게 표현할 수 있기 때문에 검색 전문가들은 불리언 질의를 사용하여 높은 검색성능을 얻는다고 알려져 있지만, 일반 사용자는 자신이 원하는 정보를 불리언 형태로 표현하는데 익숙하지 않다. 본 논문에서는 검색성능의 향상과 사용자 편의성을 동시에 만족하기 위하여 사용자의 자연어 질의를 확장 불리언 질의로 자동 변환하는 방법론을 제안한다. 먼저 자연어 질의를 범주문법에 기반한 구문분석을 수행하여 구문트리를 생성하고 연산자 및 키워드 정보를 추출하여 구문트리를 간략화한다. 다음으로 간략화된 구문트리로부터 명사구를 합성하고 키워드들에 대한 가중치를 부여한 후 불리언 질의를 생성하여 검색을 수행한다. 또한 구문분석의 오류로 인한 검색성능 저하를 최소화하기 위하여 상위 N개 구문트리에 대해 각각 불리언 질의를 생성하여 검색하는 N-BEST average 방법을 제안하였다. 정보검색 실험용 데이타 모음인 KTSET2.0으로 실험한 결과 제안된 방법은 수동으로 추출한 불리언 질의보다 8% 더 우수한 성능을 보였고, 기존의 벡터공간 모델에 기반한 자연어질의 시스템에 비해 23% 성능향상을 보였다. Abstract There have been a considerable evidence that trained users can achieve a good search effectiveness through a boolean query because a structural boolean query containing operators such as AND, OR, and NOT can make a more accurate representation of user's information need. However, it is not easy for ordinary users to construct a boolean query using appropriate boolean operators. In this paper, we propose a boolean query formulation method that automatically transforms a user's natural language query into a extended boolean query for both effectiveness and user convenience. First, a user's natural language query is syntactically analyzed using KCCG(Korean Combinatory Categorial Grammar) parser and resulting syntactic trees are structurally simplified using a tree-simplifying mechanism in order to catch the logical relationships between keywords. Next, in a simplified tree, plausible noun phrases are identified and added into the same tree as new additional keywords. Finally, a simplified syntactic tree is automatically converted into a boolean query using some mapping rules and linguistic heuristics. We also propose an N-BEST average method that uses top N syntactic trees to compensate for bad effects of single incorrect top syntactic tree. In experiments using KTSET2.0, we showed that a proposed method outperformed a traditional vector space model by 23%, and surprisingly manually constructed boolean queries by 8%.

• The Journal of the Society of Korean Medicine Diagnostics
• /
• v.10 no.2
• /
• pp.88-103
• /
• 2006

Objectives : We performed this study to search the best time for HRV measuring and which index has the validity in a short term measuring. Methods : At first, three examiner measured electrocardiogram(ECG) of 5 people(4 male + 1 female) for 20 minutes in two times. and we devided ten regions (0-5, 5-10, 10-15, 15-20, 0-10, 0-15, 0-20, 5-15, 10-20, 5-20 minutes). After that we finded the HRV index for each regions. To evaluate the validity, we used Gage Repeatability and Reproducibility Study by MINITAB(ver. 13.20). If the '%Study Var' is under 30 and 'Number of Distinct Categories' is over 4, we could tell it has validity(repeatability and reproducibility). Results : 1) The region and HRV index has repeatability and reproducibility, was HF index in 10-15 minutes. The HF index in 15-20 minutes gained 27.41(%Study Var) and 5(Number of Distinct Categories). 2) The HF index in 10-15 minutes has the better reproducibility than repeatability. The value of measurement is under the control. There is consistency among the examiner. The distribution of measurement is not large. There is no interaction between examiner and each people. 3) The HF index in 10-15, 0-10, 0-15, 0-20, 10-20 minutes was not enough to gain the repeatability and reproducibility. But it had some meanings ; %Study Var, Number of Distinct Categories is 32.68, 4 for 10-15 minutes, 34.43, 4 for 0-10 minutes, 34.44, 4 for 0-15 minutes, 33.56, 4 for 0-20 minutes, 32.07, 4 for 10-20 minutes. Conclusions : This study shows the best index of measuring HRV is the HF index in 15-20 minutes. So, from now on we should measure the HRV for 20 minutes and extract the HF index in 15-20 minutes. It must be the standard for HRV measuring.

• 대한공업교육학회지
• /
• v.43 no.1
• /
• pp.158-176
• /
• 2018

The purpose of this study is to analyze the satisfaction of the manufacturing process and the satisfaction of manufacturing lesson in the lesson of model product manufacturing based on IoT(Internet of Things) for pre-service technology teachers. and we also analyzed 'smart farm' model production manufacturing among various products. The survey questionnaires with 8 questions to investigate satisfaction level of model manufacturing process, difficulties in manufacturing, and the satisfaction level of manufacturing lesson were collected from the 15 pre-service technology teachers and analyzed by using SPSS program. and The results of this study were as follows: First, the lesson satisfaction average level of pre-service technology teachers was high(M=4.22) in model product manufacturing process for the 'smart farm' model making based on the IoT. Second, the average satisfaction level of 'patent search and prior art search report writing education' was as high as 4.07. Therefore, the application of 'prior art search report writing education' showed that it helped to make the product. Third, the best high satisfaction level in the model production manufacturing procedure was 'education of inventive thinking method'(M=4.40). Therefore, the pre-service technology teachers showed that the 'education of inventive thinking methods' was very helpful from the idea design to the optimal selection of idea. The next order of satisfaction level was high(M=4.33) in 'design education and counseling' and 'guidance through selection of professor who guide the production manufacturing in addition to professors who are in charge of lesson'. Because they were helpful in solving the lack of knowledge of pre-service technology teachers. Fourth, satisfaction level with 'the presentation of model making results and exhibitions', 'presentations and participations of external event' was high (M = 4.13). Although the results of interviews with pre-service technology teachers showed that they suffered from lack of knowledge in various technologies, but it was an opportunity to learn things and felt a sense of accomplishment.

• Journal of Intelligence and Information Systems
• /
• v.24 no.3
• /
• pp.1-19
• /
• 2018

Large amount of data is now available for research and business sectors to extract knowledge from it. This data can be in the form of unstructured data such as audio, text, and image data and can be analyzed by deep learning methodology. Deep learning is now widely used for various estimation, classification, and prediction problems. Especially, fashion business adopts deep learning techniques for apparel recognition, apparel search and retrieval engine, and automatic product recommendation. The core model of these applications is the image classification using Convolutional Neural Networks (CNN). CNN is made up of neurons which learn parameters such as weights while inputs come through and reach outputs. CNN has layer structure which is best suited for image classification as it is comprised of convolutional layer for generating feature maps, pooling layer for reducing the dimensionality of feature maps, and fully-connected layer for classifying the extracted features. However, most of the classification models have been trained using online product image, which is taken under controlled situation such as apparel image itself or professional model wearing apparel. This image may not be an effective way to train the classification model considering the situation when one might want to classify street fashion image or walking image, which is taken in uncontrolled situation and involves people's movement and unexpected pose. Therefore, we propose to train the model with runway apparel image dataset which captures mobility. This will allow the classification model to be trained with far more variable data and enhance the adaptation with diverse query image. To achieve both convergence and generalization of the model, we apply Transfer Learning on our training network. As Transfer Learning in CNN is composed of pre-training and fine-tuning stages, we divide the training step into two. First, we pre-train our architecture with large-scale dataset, ImageNet dataset, which consists of 1.2 million images with 1000 categories including animals, plants, activities, materials, instrumentations, scenes, and foods. We use GoogLeNet for our main architecture as it has achieved great accuracy with efficiency in ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Second, we fine-tune the network with our own runway image dataset. For the runway image dataset, we could not find any previously and publicly made dataset, so we collect the dataset from Google Image Search attaining 2426 images of 32 major fashion brands including Anna Molinari, Balenciaga, Balmain, Brioni, Burberry, Celine, Chanel, Chloe, Christian Dior, Cividini, Dolce and Gabbana, Emilio Pucci, Ermenegildo, Fendi, Giuliana Teso, Gucci, Issey Miyake, Kenzo, Leonard, Louis Vuitton, Marc Jacobs, Marni, Max Mara, Missoni, Moschino, Ralph Lauren, Roberto Cavalli, Sonia Rykiel, Stella McCartney, Valentino, Versace, and Yve Saint Laurent. We perform 10-folded experiments to consider the random generation of training data, and our proposed model has achieved accuracy of 67.2% on final test. Our research suggests several advantages over previous related studies as to our best knowledge, there haven't been any previous studies which trained the network for apparel image classification based on runway image dataset. We suggest the idea of training model with image capturing all the possible postures, which is denoted as mobility, by using our own runway apparel image dataset. Moreover, by applying Transfer Learning and using checkpoint and parameters provided by Tensorflow Slim, we could save time spent on training the classification model as taking 6 minutes per experiment to train the classifier. This model can be used in many business applications where the query image can be runway image, product image, or street fashion image. To be specific, runway query image can be used for mobile application service during fashion week to facilitate brand search, street style query image can be classified during fashion editorial task to classify and label the brand or style, and website query image can be processed by e-commerce multi-complex service providing item information or recommending similar item.

• Asia Marketing Journal
• /
• v.10 no.1
• /
• pp.25-64
• /
• 2008

Third-Party awards are growing in popularity. They are the hit product of the year chosen by The Korea Economic Daily, the best 10 products of the year chosen by Sports paper, the best hit product chosen by consulting firm and the best venture company of the year chosen by Information and Communication Ministry. Then these questions may be followed. Why industry likes this type of advertisement? Does this type of advertisement influences consumers' purchase intention? And if it does, how? Many researchers have been interested in external cue of product quality by focusing research effort on brand, price, producer, warranty etc. However, important but under-explored area is the role of third-party reference for signaling product quality. This paper comes from the idea that the third-party reference may signal consumers like manufacturer brand, product brand, product price, and shop brand. We develop a related theories to address research questions and drive some research hypotheses based on the previous studies probing source credibility, attribution, and signal theory. We put more emphasis on source credibility. We conducted the research based on 3x2x2x2 between group factorial design to explore causal relationship between the third party award winning advertising(real, fictional, no) and the purchase intention of consumers exposed to other information simultaneously such as product type(experience, search), distribution channel(direct, indirect) and perceived price(high, low). Since subjects are divided into 2 groups based on the means of response without extra experimental stimulus in case of perceived price. 12 different advertisements are used for conducting this study. The results are followings. First, the source credibility of the third party goes up, consumers' purchase intention would go up. It seems that consumers think the credibility of the third-party most when they are exposed to the third party award winning advertisement. Second, the product type does moderate the relationship between the third-party award winning advertisement and purchase intention. And the type of the distribution channel also moderates this relationship. The consumers' purchase intention goes up higher when they buy experience good and there is significant difference of purchase intention when consumers are exposed to direct channel treatment condition. But, perceived price has nothing to do with the third-party winning advertisement context for raising consumer intention to buy advertised product.

• Journal of Korean Academy of Nursing
• /
• v.5 no.1
• /
• pp.1-16
• /
• 1975

The newborn human is the only mammalian whose mother does not have a food supply ready for it's offspring at birth. From two to four days usually elapse before the mother's supply If milk appears, and during this period, some kind of artificial feeding should be supplied to the infants. Because of this factor, there has been continued debate fog the past hundreds of years as to when the first feeding should be started Accordingly, many experiments were carried out by scholars and because of these, Pre-lateral feedings were believed to be necessary. Many types of pre-lateral feedings were tried and the conclusion was reached that glucose water was the best food for the first infants'feedings. Traditionally, This has been started 12 hours after birth. The causes for the 12 hours delay were thought to (1) provide rest for the infants: (2) prevent regurgitation ana vomiting which tended to be prevalent during this tine: (3) in cases of low weight infants, prevention of aspiration pneumonia. From recent studies of newborn physiology and as pediatric medicine has been rapidly advancing, many studies hare been carried out concerning the improvement of infant nutrition and the early feeding of infants has been emphasized. This author believes it would be very beneficial to try two different kinds of feedings for the infant. (1) experimental feedings ana (2) comparative feeding, and during this period to investigate and compare the infants blood sugar level, hematocrit, gamma globulin level weight changes and to observe the infant reaction ill order to search for a more desirable feeding program. This study was conducted from January to March 1974 with data related to 40 healthy newborn infants (male 21, female 19: weight, 2.79∼4.20㎏ : gestation, 39∼40 weeks) born at Ewha Womens University Hospital and the results obtained were as follows : 1. At time of birth the blood sugar level from the cord sample averaged 88.99㎎/100㎖, but the blood sugar level rapidly dropped after 2 to 3 hours and reached the lowest point after 10 to 11 hours (54.48㎎/100㎖) and rose again by the 24 hour time period (76.80㎎/100㎖). Changes in the blood sugar level of the experiments: groups and the compare-five group was not significantly different until the 6 to 7 hour period, but by the 10 to 11 hour period the blood sugar levels of the experimental group (49,10㎎/100㎖) and the comparative group (49.70㎎/100㎖) were lower than the remainder of the experimental groups. 9. There ware no significant weight changes between the two groups. Average weight at birth was 3.35㎏, but at the 24 hours period birth weight averaged 3.29㎏. (1.8% reduction of birth weight). It continually lowered until at 48 hours, average weight was 3.26㎏ (2.7% reduction from birth weight.) 3. Hematocrit readings showed no significant difference between the groups. Hematocrit, the average value at birth, was 28.07% and abruptly elevated to average 64.35% at the 2 to 3 hour period, then slowly lowered to an average of 59.67% at the 6 to 7 hour period, 55.10% at the 10 to 11 hour period, ana 53.70% at the 24 hour period. 4. At birth, average gamma globulin value averaged 1,39㎎/100㎖. and at the 24 hour period averaged 1,52㎎/100㎖ revealing no significant difference between the two feeding groups. 5. Such factors as voiding, passing of meconium, regurgitation and vomiting showed no significance between the two feeding groups. However, the number of infants voiding and passing meconium in the experimental groups during the first 12 hours was slightly greater. In general there was an increased tendency for regurgitation and. vomiting among a small group of the infants during the first 24 hours which thereafter decreased. 6. Fluid intake averaged 24.38cc at the first feeding and increased to average 30.48cc at the third feeding and further increased to 73. 00cc at the fifteenth feeding. Finally it was suggested that the most reasonable method of early feeding is to give less than 25cc of 5% glucose water and/or 8% powdered milk at 8 to 9 flours after birth in order to prevent hypoglycemia and dehydration.

• Bioinformatics and Biosystems
• /
• v.1 no.2
• /
• pp.109-114
• /
• 2006

In proteomics research, proteins are digested into peptides by an enzyme and in mass spectrometer, these peptides break into fragment ions to generate tandem mass spectra. The tandem mass spectral data obtained from the mass spectrometer consists of the molecular weights of the precursor ion and fragment ions. The precursor ion mass of tandem mass spectrum is the first value that is fetched to sort the candidate peptides in the database search. We look far the peptide sequences whose molecular weight matches with precursor ion mass of the mass spectrum. Then, we choose one peptide sequence that shows the best match with fragment ions information. The precursor ion mass of the tandem mass spectrum is compared with that of the digested peptides of protein database within the mass tolerance that is assigned by users according to the mass spectrometer accuracy. In this study, we used reversed sequence database method to analyze the molecular weight distribution of precursor ions of the tandem mass spectra obtained by the FT LTQ mass spectrometer for human plasma sample. By reinterpreting the precursor ion mass distribution, we could compute the experimental accuracy and we suggested a method to improve the protein identification performance.

• Journal of Korean Academy of Nursing
• /
• v.27 no.1
• /
• pp.212-227
• /
• 1997

Cerebrovascular diseases in Korea is an important health problem since mortality and mobidity have been increased rapidly. It marked the 2nd cause of specific death rates in 1993. The stroke causes physical function disorder due to hemiparalysis and emotional disorder, and stroke patients experience helplessness, powerlessness. sense of alienation and loss of hope. These feelings make the rehabilitation difficult because they lose the will of life. The subjects of the study were seven citizens who live in Pusan, are over 50 years old and belong to low income-level. The data were collected from Jan. to Sep. 1995. The researcher as a caregiver and volunteer made confidence of them and asked for their agreement on the purpose of the study. The subjects expressed their experience as openheartedly as possible. The analysis of the data was made through the phenomenological analytic method suggested by Giorgi, which is as follows ; as an unit of description which include the subject' expressions and the researcher's observation, it is examined the theme that express the hope experience with the subject's language(underlining), and the focal meanings are identified. The focal meaning is the crystalization of the theme, which is written in the language of the researcher. After intergrating the focal meaning and make the situated structural description as the meaning of the hope experience identified on each subject's point. After intergrating the situated structural description and make the general structural description as the meaning of the hope experience identified on total subject's point then the systemizing of the structure of the hope experienced phenomena and the flowing of the conciousness was researched. The conclusions of this study was as follows : The ten sources of hope which the subjects experienced were sorted as under 〈mutual relations to others : spouse, children, relatives, fellow believer. health professioner. associate patient group〉, 〈spiritual dependence〉. 〈recovery of physical function〉. 〈rumination of the past life〉, 〈expectation of the future〉. 〈economic power〉, 〈belief〉, 〈ability〉. 〈spontaneous participation〉 and 〈recovery of roles〉. Their hope was spoken out by the following two kinds of linguistics. First. the hope was expressed in the affirmative expression as follows : 〈 to be dependable〉, 〈to make efforts〉, 〈to keep under control〉, 〈to desire〉, 〈to be pleasant〉, 〈to be peaceful〉, 〈to be grateful〉, 〈to give help〉, 〈self-confidence. Courage〉, 〈to be happy〉, 〈to satisfy oneself〉, 〈to share with others〉, 〈to understand〉 and 〈to be affected, be impressed〉 Second, the hope was expressed in the negative on pression as under : 〈to be distressed〉, 〈to be uneasy〉, 〈to be sorry, be unsatisfied〉 〈despair〉, 〈to abandon〉, 〈to be fearful〉, 〈to suffer〉, 〈to bear a burden〉 〈to be confused〉, 〈to be solitary〉, 〈chest trouble〉, 〈to feel heavy〉 〈grief〉, 〈to be daunted〉, 〈to get angry〉, 〈to be uncomfortable〉, 〈to have something regretable〉 and 〈to feel guilty〉. And their hope was expressed by the following four behavioral expressions : 〈physical sphere〉, 〈psychological sphere〉, 〈social support sphere〉 and 〈spiritual sphere〉. The reaction patterns of their hope experience appeared in the following 4 coping method : 〈conquest type〉, 〈dependence type〉, 〈adaptation type〉 and 〈fate type〉. Finally, in the hope structure the sense of certainty don't always coexict with the sense of uncertainty, When the stroke patients try to search for the best quality of life, the senses of certainty and uncertainty make a continual cyclic system in the hope structure.

• Asian Pacific Journal of Cancer Prevention
• /
• v.14 no.10
• /
• pp.5973-5981
• /
• 2013

At present, the most common cause of cancer-related death in women is breast cancer. In a large proportion of breast cancers, there is the overexpression of human epidermal growth factor receptor 2 (HER2). This receptor is a 185 KDa growth factor glycoprotein, also known as the first tumor-associated antigen for different types of breast cancers. Moreover, HER2 is an appropriate cell-surface specific antigen for passive immunotherapy, which relies on the repeated application of monoclonal antibodies that are transferred to the patient. However, vaccination is preferable because it would stimulate a patient's own immune system to actively respond to a disease. In the current study, several bioinformatics tools were used for designing synthetic peptide vaccines. PEPOP was used to predict peptides from HER2 ECD subdomain III in the form of discontinuous-continuous B-cell epitopes. Then, T-cell epitope prediction web servers MHCPred, SYFPEITHI, HLA peptide motif search, Propred, and SVMHC were used to identify class-I and II MHC peptides. In this way, PEPOP selected 12 discontinuous peptides from the 3D structure of the HER2 ECD subdomain III. Furthermore, T-cell epitope prediction analyses identified four peptides containing the segments 77 (384-391) and 99 (495-503) for both B and T-cell epitopes. This work is the only study to our knowledge focusing on design of in silico potential novel cancer peptide vaccines of the HER2 ECD subdomain III that contain epitopes for both B and T-cells. These findings based on bioinformatics analyses may be used in vaccine design and cancer therapy; saving time and minimizing the number of tests needed to select the best possible epitopes.

• Asian-Australasian Journal of Animal Sciences
• /
• v.28 no.6
• /
• pp.771-781
• /
• 2015

Thoroughbred, a relatively recent horse breed, is best known for its use in horse racing. Although myostatin (MSTN) variants have been reported to be highly associated with horse racing performance, the trait is more likely to be polygenic in nature. The purpose of this study was to identify genetic variants strongly associated with racing performance by using estimated breeding value (EBV) for race time as a phenotype. We conducted a two-stage genome-wide association study to search for genetic variants associated with the EBV. In the first stage of genome-wide association study, a relatively large number of markers (~54,000 single-nucleotide polymorphisms, SNPs) were evaluated in a small number of samples (240 horses). In the second stage, a relatively small number of markers identified to have large effects (170 SNPs) were evaluated in a much larger number of samples (1,156 horses). We also validated the SNPs related to MSTN known to have large effects on racing performance and found significant associations in the stage two analysis, but not in stage one. We identified 28 significant SNPs related to 17 genes. Among these, six genes have a function related to myogenesis and five genes are involved in muscle maintenance. To our knowledge, these genes are newly reported for the genetic association with racing performance of Thoroughbreds. It complements a recent horse genome-wide association studies of racing performance that identified other SNPs and genes as the most significant variants. These results will help to expand our knowledge of the polygenic nature of racing performance in Thoroughbreds.