• Journal of Intelligence and Information Systems
• /
• v.25 no.4
• /
• pp.105-122
• /
• 2019

Dimensionality reduction is one of the methods to handle big data in text mining. For dimensionality reduction, we should consider the density of data, which has a significant influence on the performance of sentence classification. It requires lots of computations for data of higher dimensions. Eventually, it can cause lots of computational cost and overfitting in the model. Thus, the dimension reduction process is necessary to improve the performance of the model. Diverse methods have been proposed from only lessening the noise of data like misspelling or informal text to including semantic and syntactic information. On top of it, the expression and selection of the text features have impacts on the performance of the classifier for sentence classification, which is one of the fields of Natural Language Processing. The common goal of dimension reduction is to find latent space that is representative of raw data from observation space. Existing methods utilize various algorithms for dimensionality reduction, such as feature extraction and feature selection. In addition to these algorithms, word embeddings, learning low-dimensional vector space representations of words, that can capture semantic and syntactic information from data are also utilized. For improving performance, recent studies have suggested methods that the word dictionary is modified according to the positive and negative score of pre-defined words. The basic idea of this study is that similar words have similar vector representations. Once the feature selection algorithm selects the words that are not important, we thought the words that are similar to the selected words also have no impacts on sentence classification. This study proposes two ways to achieve more accurate classification that conduct selective word elimination under specific regulations and construct word embedding based on Word2Vec embedding. To select words having low importance from the text, we use information gain algorithm to measure the importance and cosine similarity to search for similar words. First, we eliminate words that have comparatively low information gain values from the raw text and form word embedding. Second, we select words additionally that are similar to the words that have a low level of information gain values and make word embedding. In the end, these filtered text and word embedding apply to the deep learning models; Convolutional Neural Network and Attention-Based Bidirectional LSTM. This study uses customer reviews on Kindle in Amazon.com, IMDB, and Yelp as datasets, and classify each data using the deep learning models. The reviews got more than five helpful votes, and the ratio of helpful votes was over 70% classified as helpful reviews. Also, Yelp only shows the number of helpful votes. We extracted 100,000 reviews which got more than five helpful votes using a random sampling method among 750,000 reviews. The minimal preprocessing was executed to each dataset, such as removing numbers and special characters from text data. To evaluate the proposed methods, we compared the performances of Word2Vec and GloVe word embeddings, which used all the words. We showed that one of the proposed methods is better than the embeddings with all the words. By removing unimportant words, we can get better performance. However, if we removed too many words, it showed that the performance was lowered. For future research, it is required to consider diverse ways of preprocessing and the in-depth analysis for the co-occurrence of words to measure similarity values among words. Also, we only applied the proposed method with Word2Vec. Other embedding methods such as GloVe, fastText, ELMo can be applied with the proposed methods, and it is possible to identify the possible combinations between word embedding methods and elimination methods.

• Annals of Clinical Microbiology
• /
• v.21 no.4
• /
• pp.86-91
• /
• 2018

Background: The PANA RealTyper HPV kit (PANAGENE, Korea; PANA RealTyper) was developed to genotype human papillomavirus (HPV) and was based on multiplex real-time PCR amplification and melting curve analysis. In this study, we compared PANA RealTyper to the AdvanSure HPV GenoBlot assay (LG Life Sciences, Korea; AdvanSure assay) and attempted to evaluate the performance of PANA RealTyper. Methods: A total of 60 cervical specimens were collected from women undergoing routine cervical cancer screening. The AdvanSure assay and PANA RealTyper kit identified the same 20 high-risk genotypes. However, the AdvanSure assay identified 15 low-risk genotypes, while the PANA RealTyper kit identified only 2 but detected 18 low-risk genotypes. Results: Among the total 60 specimens, 54 high-risk genotypes (40 specimens) and 20 low-risk genotypes (18 specimens) were detected. The agreement rates of the assays ranged from 94.4 to 100% for high-risk genotypes. Among 9 genotypes that were positive in the PANA RealTyper kit but negative in the AdvanSure assay, 7 were confirmed as true positive (HPV genotypes 16 (n=1), 39 (n=1), 52 (n=1), 58 (n=2), 68 (n=2)). Among 4 genotypes that were negative in the PANA RealTyper kit but positive in the AdvanSure assay, 3 were confirmed as HPV genotype 59. Among the 19 low-risk genotypes positive in the AdvanSure assay, there were 2 cases of HPV 6 and 1 case of HPV 11. In comparison, only 1 positive case of HPV 6 was determined by the PANA RealTyper kit. Conclusion: The PANA RealTyper kit was comparable with the AdvanSure assay. The PANA RealTyper kit would be useful and suitable for HPV genotyping in the clinical laboratory.

• Journal of IKEEE
• /
• v.25 no.3
• /
• pp.493-500
• /
• 2021

In this paper, we propose the development of deep learning structure to improve quality of polygonal containers. The deep learning structure consists of a convolution layer, a bottleneck layer, a fully connect layer, and a softmax layer. The convolution layer is a layer that obtains a feature image by performing a convolution 3x3 operation on the input image or the feature image of the previous layer with several feature filters. The bottleneck layer selects only the optimal features among the features on the feature image extracted through the convolution layer, reduces the channel to a convolution 1x1 ReLU, and performs a convolution 3x3 ReLU. The global average pooling operation performed after going through the bottleneck layer reduces the size of the feature image by selecting only the optimal features among the features of the feature image extracted through the convolution layer. The fully connect layer outputs the output data through 6 fully connect layers. The softmax layer multiplies and multiplies the value between the value of the input layer node and the target node to be calculated, and converts it into a value between 0 and 1 through an activation function. After the learning is completed, the recognition process classifies non-circular glass bottles by performing image acquisition using a camera, measuring position detection, and non-circular glass bottle classification using deep learning as in the learning process. In order to evaluate the performance of the deep learning structure to improve quality of polygonal containers, as a result of an experiment at an authorized testing institute, it was calculated to be at the same level as the world's highest level with 99% good/defective discrimination accuracy. Inspection time averaged 1.7 seconds, which was calculated within the operating time standards of production processes using non-circular machine vision systems. Therefore, the effectiveness of the performance of the deep learning structure to improve quality of polygonal containers proposed in this paper was proven.

• Korean Journal of Environmental Biology
• /
• v.36 no.4
• /
• pp.517-524
• /
• 2018

The purpose of this study was to investigate the density of thrips, and insecticidal resistance for effective control of Western flower thrips in greenhouse. The presence and density of the thrips was investigated using yellow colored-sticky trap in a cucumber field from May to August in Cheon-an. The results of the investigation revealed the existence of the following thrips species; Frankliniella occidentalis, F. intonsa, Thrips palmi, T. tabaci, Scirtothrips dorsalis, Microcephalothrips abdominalis, and T. nigropilosus. The predominant pest was found to be the western flower thrips. To survey the western flower thrips insecticidal resistance, we established the discriminating concentration (DC), recommended concentration (RC) and $2{\times}$recommended concentration ($2{\times}RC$) of nine insecticides; Emamectin benzoate EC, spinetoram SC, Chlorfenapyr EC, Spinosad SC, Cyantraniliprole EC, Acetamiprid WP, Dinotefuran WG, thiacloprid SC and thiamethoxam SC. The bioassay of about five local populations was conducted using the leaf-dipping method. In all local populations, insecticidal resistance in western flower thrips had not developed in emamectin benzoate EC (RC, $10.8{\mu}L\;L^{-1}$), chlorfenapyr EC (RC, $50.0{\mu}L\;L^{-1}$), spinetoram SC (RC, $25.0mg\;L^{-1}$), and spinosad SC (RC, $50.0mg\;L^{-1}$). However, insecticidal resistance in RC was found to have developed in cyantraniliprole EC (RC, $50.0{\mu}L\;L^{-1}$) and four insecticides of neonicotinoid type. Insecticidal activity of 95% or more was observed in each population when cyantraniliprole EC tested in $2{\times}RC$. However, the neonicotinoid types showed different insecticidal activity in $2{\times}RC$.

• The Journal of the Petrological Society of Korea
• /
• v.28 no.1
• /
• pp.1-13
• /
• 2019

The characteristics of the rock cleavage of Jurassic Geochang granite were analysed using the parameters from the length and spacing-cumulative frequency diagrams. The evaluation for three planes and three rock cleavages was performed using the 25 parameters such as (1~2) slope angle(${\alpha}^{\circ}$and ${\beta}^{\circ}$), (3) intersection angle(${\alpha}-{\beta}^{\circ}$), (4) exponent difference(${\lambda}_S-{\lambda}_L$), (5~12) length of line(oa, ob, ol, os, ss', ll' and sl') and (13~15) length ratio(ol/os, ss'/ll' and ll'/sl'), (16) mean length((ss'+ll')/2), (17~23) area (${\Delta}oaa^{\prime}$, ${\Delta}obb^{\prime}$, ${\Delta}obb^{\prime}$, ${\Delta}oaa_a^{\prime}$, ${\Delta}obb_a^{\prime}$, ${\Delta}ll^{\prime}s^{\prime}$, ${\Delta}ss^{\prime}l^{\prime}$ and ⏢$ll^{\prime}ss^{\prime}$) and (24~25) area difference(${\Delta}obb^{\prime}-{\Delta}oaa^{\prime}$ and ${\Delta}obb_a^{\prime}-{\Delta}oaa_a^{\prime}$). Firstly, the values of the 11 parameters(group I: No. 1, 3~4, 7, 9~10, 13, 15~16, 20 and 25), the 3 parameters(group II: No. 5, 8 and 17) and the 2 parameters(group III: No. 12 and 22) are in orders of H(hardway) < G(grain) < R(rift), R < G < H and G < H < R, respectively. On the contrary, the values of parameters belonging to the above three groups show reverse orders for three planes. Secondly, the generalized chart for three planes and three rock cleavages were made. From the related chart, the distribution types formed by the two diagrams related to lengths and spacings were derived. The diagrams related to spacings show upward curvature in the chart of rift plane(G1 & H1, R') and hardway(H1 & H2, H). On the contrary, the diagrams related to lengths show downward curvature. These two diagrams take the form of a convex lens in the upper section. Besides, the two diagrams cross each other in the lower section. The overall shape formed by the above two diagrams between three planes($H^{\prime}{\rightarrow}G^{\prime}{\rightarrow}R^{\prime}$) and three rock cleavages($R{\rightarrow}G{\rightarrow}H$) display in reverse order. Lastly, these types of correlation analysis is useful for discriminating three quarrying planes.

• The Journal of Korean Academy of Sensory Integration
• /
• v.19 no.1
• /
• pp.69-82
• /
• 2021

Objective : This recent work intended to provide basic information for researchers and practitioners related to occupational therapy about Developmental Coordination Disorder (DCD) in South Korea. The previous research of screening DCD and the effects of intervention programs were reviewed. Methods : Peer-reviewed papers relating to DCD and published in Korea from January 1990 to December 2020 were systematically reviewed. The search terms "developmental coordination disorder," "development coordination," and "developmental coordination" were used to identify previous Korean research in this area from three representation database, the Research Information Sharing Service, Korean Studies Information Service System, and Google Scholar. We found a total of 4,878 articles identified through the three search engines and selected seventeen articles for analysis after removing those that corresponded to the overlapping or exclusion criteria. We adopted "the conceptual model" to analyze the selected articles about DCD assessment and intervention. Results : We found that twelve of the 17 studies showed the qualitative level of Level 2 using non-randomized approach between the two groups. The Movement Assessment Battery for Children and its second edition were the most frequently used tools in assessing children for DCD. Among the intervention studies, the eight articles (47%) were adopted a dynamic systems approach; a normative functional skill framework and cognitive neuroscience were each used in 18% of the pieces; and 11% of the articles were applied neurodevelopmental theory. Only one article was used a combination approach of normative functional skill and general abilities. These papers were mainly focused on the movement characteristics of children with DCD and the intervention effect of exercise or sports programs. Conclusion : Most of the reviewed studies investigated the movement characteristics of DCD or explore the effectiveness of particular intervention programs. In the future, it would be useful to investigate the feasibility of different assessment tools and to establish the effectiveness of various interventions used in rehabilitation for better motor performance in children with DCD.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.51 no.2
• /
• pp.94-102
• /
• 2023

The Florence Charter, established in 1981 by the ICOMOS-IFLA International Commission on Historic Gardens, considered the historic garden as a plant-dominated architectural composition, a perpetual balance between the artist and the artisan's desire to remain unaltered in perpetuity. Trees, the main component of the garden, require continuous management as they repeat their creation and calling according to the cycle of the seasons and accordingly, it is suggested that management to maintain the appearance of trees in Korean palaces was an indispensable element. Although it is an important matter to historically verify the tree management techniques of the palaces of the past, research has been difficult due to the absence of historical records and the disconnection of life due to the Japanese colonial period. In addition, according to the view that trees were not generally managed in the palace, research on palace tree management techniques has not been conducted so far. Therefore, this study aimed to examine whether or not the tree management of the palaces was performed in the past through expert interviews based on photographs taken in modern times. With the identification of tree species and pruning through in-depth interviews with experts using photographs taken in the modern period, the results are as follows. First, it was found that the shapes of trees and leaves could be identified through photographs and by observing the phenomena caused by planting and pruning, it was possible to estimate whether or not the tree management was implemented in the modern palace. Second, as a result of in-depth interviews with 8 experts in 4 fields, it was possible to determine the tree species and pruning status, purpose, and method. There was no significant difference in opinions between groups, and the evidence was clearly presented. Third, the type of management of trees in the palace was mainly found to be determination of the types of tree, removal of hazardous factors of trees, and management of lower vegetation and through the photographs of Jondeokjeong and Gwanwanjeong, it was confirmed that the trees were managed even before the Japanese colonial era, considering the time of filming. Based on the photographs taken, it was possible to estimate whether the trees were managed before the Japanese colonial period through expert interviews. However, it could not be clarified due to lack of historical materials as to whether it was carried out on its own according to the circumstances of the times or by the Japanese Empire. Still, in this study, evidence was obtained to refute the view that trees of the palace were not managed in the past through the collected data. Expert opinions supporting this view were collected to make the conclusion. In addition, based on the general theory of pruning, an empirical review of expert opinions was conducted to secure the reliability of the research results.

• Journal of the Korean Society of Radiology
• /
• v.17 no.3
• /
• pp.385-392
• /
• 2023

Caroli's disease is a fibrocystic liver disease. Autosomal recessive disorder is characterized by congenital multiple dilatation of the bile duct. Computerized tomography, magnetic resonance imaging, cholangiography and ultrasound are among the methods for diagnosing caroli disease. Computerized tomography is essential for detecting and distinguishing fibroplastic liver disease and is useful for determining intrahepatic bile duct dilatation. However, awareness of the possible side effects of using contrast mediums is necessary. A typical method of magnetic resonance cholangiography is used for magnetic resonance imaging. A non-invasive examination can reduce the pain of the patient, and the anatomical structure of the bile pancreatic duct and the presence or absence of lesions can be easily and quickly observed. Biliary contrast is an effective diagnostic method that can directly visualize various cystic dilatations throughout the enlarged bile duct. However, since this procedure is also an invasive procedure, it is recommended not for diagnosis but for treatment purposes. Ultrasonography can confirm similar findings to computerized tomography. The hepatic artery root is difficult to prove with conventional grayscale ultrasound. However, it is of clinical value in that it can not only describe dilated bile ducts with vascular roots in the tube but also easily identify color Doppler signals in the tube. With the development of video diagnostics, early diagnosis has become possible through computerized tomography, magnetic resonance imaging, cholangiography, and ultrasound. In order to further contribute to the development of video diagnostics so that long-term prognosis can be improved after treatment through early diagnosis, we examined what aspects of each test's caroli disease appear.

• Economic and Environmental Geology
• /
• v.51 no.3
• /
• pp.213-222
• /
• 2018

Metasediment-hosted Pb-Zn mineralized zone has been found in Dyusembay of Kazakhstan. Its petrological properties, metal index, alteration index and redox-sensitivity are compared with those of SEDEX type deposit. Mineralization is developed along foliation of host rock (graphitic phyllite) and controlled by folds and faults; major ore minerals including pyrite, pyrrhotite, sphalerite, and galena are disseminated or interlayered with fine-grained quartz. The margin of the mineralized zone is metamorphosed accompanying sericite and chlorite. Hydrothermal brecciation and Pb-Zn mineralization formed in quartz-calcite stockworks are confirmed at the around of Maytyubin granitoid intrusions. The mineralization is classified into three types according to those of occurrence, paragenesis, chemical composition and isotopic characteristics. Type 1 whose fine-grained pyrite, pyrrhotite and sphalerite are formed in parallel yet discontinuous to well-developed foliations of the host rock; its geochemistry is similar to those of the earlier stage in SEDEX-type mineralization. In case of type 2, the ore minerals of which are concentrated being parallel to a foliation by regional metamorphism, and most of them associated with quartz and muscovite (${\pm}$ biotite) paragenetically. Type 3 is formed in the hydrothermal breccia zone whose ore minerals are controlled by foliation and breccia and developed in quartz ${\pm}$ calcite veins having a form such as stratification, stockwork or veinlets. Host rocks in the mineralized zone indicate homogeneous metamorphic grade and there is no specific alteration zonation. Also, all types (type 1, type 2, and type 3) represent similar REEs patterns, it can be interpreted that these are originated from a same source. Sulphides occurred in mineralized zone indicate a limited range of sulphur isotope values (type 2, ${\delta}^{34}S=-13.3{\sim}-11.7$‰; type 3, ${\delta}^{34}S=-13.9{\sim}-8.2$‰), and a result of geothermometry presents different temperature ranges: type 2($251{\pm}38^{\circ}C{\sim}277{\pm}40^{\circ}C$); type 3($360{\pm}2^{\circ}C$ to $537{\pm}29^{\circ}C$). It is estimated to be due to the effect of metamorphism and Maytyubin granitoid intrusions, respectively. In addition, ternary chart of thorium, scandium, and zircon for discrimination of tectonic setting and redox sensitivity using V/Mo values indicate that hydrothermal sediments put on reduction environment after precipitation, before being affected by metamorphism and intrusion activity. Geochemical data are plotted on a distal trend of SEDEX-type with discrimination plot using SEDEX index. As a result, petrological-geochemical properties demonstrate that Dyusembay Pb-Zn mineralized zone is comparable to distal-type of SEDEX deposit.

• Journal of Intelligence and Information Systems
• /
• v.19 no.3
• /
• pp.113-125
• /
• 2013

Every company wants to know customer's requirement and makes an effort to meet them. Cause that, communication between customer and company became core competition of business and that important is increasing continuously. There are several strategies to find customer's needs, but VOC (Voice of customer) is one of most powerful communication tools and VOC gathering by several channels as telephone, post, e-mail, website and so on is so meaningful. So, almost company is gathering VOC and operating VOC system. VOC is important not only to business organization but also public organization such as government, education institute, and medical center that should drive up public service quality and customer satisfaction. Accordingly, they make a VOC gathering and analyzing System and then use for making a new product and service, and upgrade. In recent years, innovations in internet and ICT have made diverse channels such as SNS, mobile, website and call-center to collect VOC data. Although a lot of VOC data is collected through diverse channel, the proper utilization is still difficult. It is because the VOC data is made of very emotional contents by voice or text of informal style and the volume of the VOC data are so big. These unstructured big data make a difficult to store and analyze for use by human. So that, the organization need to automatic collecting, storing, classifying and analyzing system for unstructured big VOC data. This study propose an intelligent VOC analyzing system based on opinion mining to classify the unstructured VOC data automatically and determine the polarity as well as the type of VOC. And then, the basis of the VOC opinion analyzing system, called domain-oriented sentiment dictionary is created and corresponding stages are presented in detail. The experiment is conducted with 4,300 VOC data collected from a medical website to measure the effectiveness of the proposed system and utilized them to develop the sensitive data dictionary by determining the special sentiment vocabulary and their polarity value in a medical domain. Through the experiment, it comes out that positive terms such as "칭찬, 친절함, 감사, 무사히, 잘해, 감동, 미소" have high positive opinion value, and negative terms such as "퉁명, 뭡니까, 말하더군요, 무시하는" have strong negative opinion. These terms are in general use and the experiment result seems to be a high probability of opinion polarity. Furthermore, the accuracy of proposed VOC classification model has been compared and the highest classification accuracy of 77.8% is conformed at threshold with -0.50 of opinion classification of VOC. Through the proposed intelligent VOC analyzing system, the real time opinion classification and response priority of VOC can be predicted. Ultimately the positive effectiveness is expected to catch the customer complains at early stage and deal with it quickly with the lower number of staff to operate the VOC system. It can be made available human resource and time of customer service part. Above all, this study is new try to automatic analyzing the unstructured VOC data using opinion mining, and shows that the system could be used as variable to classify the positive or negative polarity of VOC opinion. It is expected to suggest practical framework of the VOC analysis to diverse use and the model can be used as real VOC analyzing system if it is implemented as system. Despite experiment results and expectation, this study has several limits. First of all, the sample data is only collected from a hospital web-site. It means that the sentimental dictionary made by sample data can be lean too much towards on that hospital and web-site. Therefore, next research has to take several channels such as call-center and SNS, and other domain like government, financial company, and education institute.