• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.175-197
• /
• 2021

Recently, with the development of deep learning technology, research on unstructured data analysis is being actively conducted, and it is showing remarkable results in various fields such as classification, summary, and generation. Among various text analysis fields, text classification is the most widely used technology in academia and industry. Text classification includes binary class classification with one label among two classes, multi-class classification with one label among several classes, and multi-label classification with multiple labels among several classes. In particular, multi-label classification requires a different training method from binary class classification and multi-class classification because of the characteristic of having multiple labels. In addition, since the number of labels to be predicted increases as the number of labels and classes increases, there is a limitation in that performance improvement is difficult due to an increase in prediction difficulty. To overcome these limitations, (i) compressing the initially given high-dimensional label space into a low-dimensional latent label space, (ii) after performing training to predict the compressed label, (iii) restoring the predicted label to the high-dimensional original label space, research on label embedding is being actively conducted. Typical label embedding techniques include Principal Label Space Transformation (PLST), Multi-Label Classification via Boolean Matrix Decomposition (MLC-BMaD), and Bayesian Multi-Label Compressed Sensing (BML-CS). However, since these techniques consider only the linear relationship between labels or compress the labels by random transformation, it is difficult to understand the non-linear relationship between labels, so there is a limitation in that it is not possible to create a latent label space sufficiently containing the information of the original label. Recently, there have been increasing attempts to improve performance by applying deep learning technology to label embedding. Label embedding using an autoencoder, a deep learning model that is effective for data compression and restoration, is representative. However, the traditional autoencoder-based label embedding has a limitation in that a large amount of information loss occurs when compressing a high-dimensional label space having a myriad of classes into a low-dimensional latent label space. This can be found in the gradient loss problem that occurs in the backpropagation process of learning. To solve this problem, skip connection was devised, and by adding the input of the layer to the output to prevent gradient loss during backpropagation, efficient learning is possible even when the layer is deep. Skip connection is mainly used for image feature extraction in convolutional neural networks, but studies using skip connection in autoencoder or label embedding process are still lacking. Therefore, in this study, we propose an autoencoder-based label embedding methodology in which skip connections are added to each of the encoder and decoder to form a low-dimensional latent label space that reflects the information of the high-dimensional label space well. In addition, the proposed methodology was applied to actual paper keywords to derive the high-dimensional keyword label space and the low-dimensional latent label space. Using this, we conducted an experiment to predict the compressed keyword vector existing in the latent label space from the paper abstract and to evaluate the multi-label classification by restoring the predicted keyword vector back to the original label space. As a result, the accuracy, precision, recall, and F1 score used as performance indicators showed far superior performance in multi-label classification based on the proposed methodology compared to traditional multi-label classification methods. This can be seen that the low-dimensional latent label space derived through the proposed methodology well reflected the information of the high-dimensional label space, which ultimately led to the improvement of the performance of the multi-label classification itself. In addition, the utility of the proposed methodology was identified by comparing the performance of the proposed methodology according to the domain characteristics and the number of dimensions of the latent label space.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.24 no.2
• /
• pp.12-28
• /
• 2021

Based on the global smart city promotion trend, in 2018, the "Fourth Industrial Revolution Committee" selected "sustainability" and "people-centered" as keywords in relation to the direction of domestic smart city policy. Accordingly, the Living Lab program, which is an active citizen-centered innovation methodology, is applied to each stage of the domestic smart city construction project. Through the Living Lab program, and in collaboration with the public and experts, the smart city discovers local issues as it focuses on citizens, devises solutions to sustainable urban problems, and formulates a regional development plan that reflects the needs of citizens. However, compared to citizen participation in urban regeneration projects that have been operated for a relatively long time, participation in smart city projects was found to significantly differ in level and sustainability. Therefore, this study conducted a comparative analysis of the characteristics of citizen participation at each stage of an urban regeneration project and, based on Arnstein's "Participation Ladder" model, examined the level of citizen participation activities in the Living Lab program carried out in a smart city commercial area from 2018 to 2019. The results indicated that citizen participation activities in the Living Lab conducted in the smart city project had a great influence on selecting smart city services, which fit the needs of local residents, and on determining the technological level of services appropriate to the region based on a relatively high level of authority, such as selection of smart city services or composition of solutions. However, most of the citizen participation activities were halted after the project's completion due to the one-off recruitment of citizen participation groups for the smart city construction project only. On the other hand, citizens' participation activities in the field of urban regeneration were focused on local communities, and continuous operation and management measures were being drawn from the project planning stage to the operation stage after the project was completed. This study presented a plan to revitalize citizen participation for the realization of a more sustainable smart city through a comparison of the characteristics and an examination of the level of citizen participation in such urban regeneration and smart city projects.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.23 no.4
• /
• pp.374-390
• /
• 2021

Growth and yield of field vegetable crops would be affected by climate conditions, which cause a relatively large fluctuation in crop production and consumer price over years. The yield prediction system for these crops would support decision-making on policies to manage supply and demands. The objectives of this study were to compile literatures related to onion and garlic and to perform data-mining analysis, which would shed lights on the development of crop models for these major field vegetable crops in Korea. The literatures on crop growth and yield were collected from the databases operated by Research Information Sharing Service, National Science & Technology Information Service and SCOPUS. The keywords were chosen to retrieve research outcomes related to crop growth and yield of onion and garlic. These literatures were analyzed using text mining approaches including word cloud and semantic networks. It was found that the number of publications was considerably less for the field vegetable crops compared with rice. Still, specific patterns between previous research outcomes were identified using the text mining methods. For example, climate change and remote sensing were major topics of interest for growth and yield of onion and garlic. The impact of temperature and irrigation on crop growth was also assessed in the previous studies. It was also found that yield of onion and garlic would be affected by both environment and crop management conditions including sowing time, variety, seed treatment method, irrigation interval, fertilization amount and fertilizer composition. For meteorological conditions, temperature, precipitation, solar radiation and humidity were found to be the major factors in the literatures. These indicate that crop models need to take into account both environmental and crop management practices for reliable prediction of crop yield.

• Journal of the Daesoon Academy of Sciences
• /
• v.37
• /
• pp.191-227
• /
• 2021

In this study, values of truth and good are expressed in the form of beauty, and truth and good are analyzed from an aesthetic point of view. This enables an assessment of how truth is expressed and presented as an "aesthetic" in Daesoon Thought. Therefore, an approach to faith in Daesoon Jinrihoe (大巡眞理會) can be presented via traditional aesthetics or theological aesthetics that reflect on sense experience, feelings, and beauty. The concept of beauty in Daesoon Thought which focuses on The Canonical Scripture appears in keywords used in Daesoon Thought such as divine nature (神性), the pattern of Dao (道理), the singularly-focused mind (一心), and relationships (關係). Therein, one can find sublimation, symmetry, moderation, and harmony. The aesthetic features of Daesoon Thought, when considered as an aesthetic system can formulate thinking regarding the aesthetics of 'Reordering Works of Heaven and Earth' (天地公事), the aesthetics of Mutual Beneficence (相生), and the aesthetics of healing. The Reordering Works of Heaven and Earth contain a record of the Supreme God visiting the world as a human being. The realization that the human figure, Kang Jeungsan (1871-1909), is the Supreme God, Sangje (上帝), is the shocking aesthetic motif and theological starting point of the Reordering Works of Heaven and Earth. Mutual Beneficence can be seen aesthetically as indicating the sociality of mutual relations, and there is an aesthetic structure of Mutual Beneficence in the harmony and unification of those relations. Healing can be said to contain the sacred sublimation of Sangje, and moderation is a form of beauty that makes humans move toward Quieting the mind and Quieting the body (安心·安身), the Dharma of Presiding over Cures (醫統), and the ultimate value of healing, which is the end point of the Cultivation (修道) wherein one realizes that the ideals of humankind and the aesthetics of healing bestow the spiritual pleasures of a beautiful and valuable life. The aesthetic characteristics of Daesoon Thought demonstrate an aesthetic attitude that leads to healing through Sangje's Holy Works and the practice of Mutual Beneficence (相生) which were performed when He stayed with us to vastly save all beings throughout the Three Realms that teetered on the brink of extinction. It is not uncommon to see a beautiful woman and remark she is like a goddess (女神) or female immortal (仙女). Likewise, beautiful music is often praised as "the sound of heaven." That which fills us with joy is spoken of as "divine beings (神明)" of God. God is a symbol of beauty, and the world of God can be said to be the archetype of beauty. Experience of beauty guides our souls to God. The aesthetic experience of Daesoon Thought is a religious experience that culminates in emotional, intellectual, and spiritual joy, and it is an aesthetic experience that recognizes transcendent beauty.

• Journal of the Society of Disaster Information
• /
• v.18 no.1
• /
• pp.202-217
• /
• 2022

Purpose: The recent increased use of Personal Mobility (PM) has been accompanied by a rise in the annual number of accidents. Accordingly, the safety requirements for PM use are being strengthened, but the laws/systems, infrastructure, and management systems remain insufficient for fostering a safe environment. Therefore, this study comprehensively searches the main problems and improvement methods through a review of previous studies that are related to PM. Then the priorities according to the importance of the improvement methods are presented through the Delphi survey. Method: The research method is mainly composed of a literature study and an expert survey (Delphi survey). Prior research and improvement cases (local governments, government departments, companies, etc.) are reviewed to derive problems and improvements, and a problem/improvement classification table is created based on keywords. Based on the classification contents, an expert survey is conducted to derive a priority improvement plan. Result: The PM-related problems were in 'non-compliance with traffic laws, lack of knowledge, inexperienced operation, and lack of safety awareness' in relation to human factors, and 'device characteristics, road-drivable space, road facilities, parking facilities' in relation to physical factors. 'Management/supervision, product management, user management, education/training' as administrative factors and legal factors are divided into 'absence/sufficiency of law, confusion/duplication, reduced effectiveness'. Improvement tasks related to this include 'PM education/public relations, parking/return, road improvement, PM registration/management, insurance, safety standards, traffic standards, PM device safety, PM supplementary facilities, enforcement/management, dedicated organization, service providers, management system, and related laws/institutional improvement', and 42 detailed tasks are derived for these 14 core tasks. The results for the importance evaluation of detailed tasks show that the tasks with a high overall average for the evaluation items of cost, time, effect, urgency, and feasibility were 'strengthening crackdown/instruction activities, education publicity/campaign, truancy PM management, and clarification of traffic rules'. Conclusion: The PM market is experiencing gradual growth based on shared services and a safe environment for PM use must be ensured along with industrial revitalization. In this respect, this study seeks out the major problems and improvement plans related to PM from a comprehensive point of view and prioritizes the necessary improvement measures. Therefore, it can serve as a basis of data for future policy establishment. In the future, in-depth data supplementation will be required for each key improvement area for practical policy application.

• 한국노년학
• /
• v.40 no.4
• /
• pp.801-821
• /
• 2020

There are a number of prior studies on the poverty experience of Korean women, but little is known about the poverty experience of Korean elderly women in the U.S. The purpose of this study is to examine the poverty experiences of Korean elderly women who immigrated to the U. S. Qualitative case study methods were used to achieve these research objectives. Three Korean elderly women living in Oakland of California who received Supplemental Security Income (SSI) from the U.S. federal government were included in the study. The data were collected by conducting a total of six meetings per participant, and the researcher read the consent form directly to the participants and obtained a hand-written signature. The analysis and interpretation began by repeating the interview transcript several times, and the repeated keywords were to be understood in the context, focusing on time, space, and relationships with other people. The contextual understanding of Korean elderly women's experiences in poverty was interpreted in three dimensions: extending poverty in their mother country, double torture as female immigrants, and limiting labor due to aging and diseases. Before moving to the U.S., they had a difficult livelihood by farming and one of them had to live in poverty due to the bereavement to her husband. But even after moving to the U.S., they have continued to live in poverty. As female immigrants with low education and no special skills, they were incorporated into the periphery of the labor market in the industrialized U.S. and were forced to make a living with low wages. Korean elderly women were unable to return to the labor market in the surrounding areas due to aging and diseases, and were continuing their impoverished lives relying on SSI. From the findings, we discussed the role of the Korean immigrants community as a way to improve the quality of life for Korean elderly women in the U.S.

• Journal of Christian Education in Korea
• /
• v.71
• /
• pp.181-210
• /
• 2022

The purpose of this study was to investigate the research trends related to the children's department of church schools. The purpose of this study is to present basic data for the study of the children's department of church schools by analyzing the research period, research contents, research methods, and subjects of research related to the children's department of church schools. For this study, 50 domestic master's and doctoral dissertations searched through the National Assembly Library and the Research Information Sharing Service(RISS) were extracted with the keywords of 'church school' and 'children's department'. The frequency and percentage were calculated by analyzing the research related to the children's department of the church school according to four criteria: research period, research content, research method, and research subject. As a result of the study, first, the research trend of research papers in the children's department of church schools was found to be 49 articles (98%) for master's degrees and 1 article (2%) for doctoral degrees from 1980 to 2022. Trends by research period are focused on master's degrees. Second, the trend by research content was 27 practical studies (54%) and 23 theory studies (46%). In the research related to the children's department of church schools, the practical research accounted for a relatively high percentage compared to the theory research. Third, the trends by research method were in the order of 30 literature studies (60%), 19 quantitative studies (38%), and 1 qualitative study (2%). Research related to children's departments in church schools is being actively conducted with a focus on literature research. Fourth, as for the trends by study subject, the study was conducted focusing on physical subjects, with 35 subjects (70%) and 15 subjects (30%) of personal subjects. As research is conducted from physical objects to church schools and media, it is necessary to study the connection between church schools and families. As the research on church school kindergarten is focused on adults (teachers, parents, and educational preachers), in-depth research on children in church schools and qualitative research with voices from the field of children's department in church schools are required.

• Journal of Korea Entertainment Industry Association
• /
• v.14 no.6
• /
• pp.139-150
• /
• 2020

The data processing of this study focuses on the textom and social media words about three areas: 'Corona 19 and professional baseball', 'Corona 19 and professional baseball', and 'Corona 19 and professional sports' The data was collected and refined in a web environment and then processed in batch, and the Ucinet6 program was used to visualize it. Specifically, the web environment was collected using Naver, Daum, and Google's channels, and was summarized into 30 words through expert meetings among the extracted words and used in the final study. 30 extracted words were visualized through a matrix, and a CONCOR analysis was performed to identify clusters of similarity and commonality of words. As a result of analysis, the clusters related to Corona 19 and Pro Baseball were composed of one central cluster and five peripheral clusters, and it was found that the contents related to the opening of professional baseball according to the corona 19 wave were mainly searched. The cluster related to Corona 19 and unrelated to professional baseball consisted of one central cluster and five peripheral clusters, and it was found that the keyword of the position of professional baseball related to the professional baseball game according to Corona 19 was mainly searched. Corona 19 and the cluster related to professional sports consisted of one central cluster and five peripheral clusters, and it was found that the keywords related to the start of professional sports according to the aftermath of Corona 19 were mainly searched.

• Journal of Christian Education in Korea
• /
• v.68
• /
• pp.245-277
• /
• 2021

The purpose of this study was to investigate the trend of Christian picture book-related research. The purpose of this study is to present basic data for various and balanced research and development in the Christian picture book field by analyzing the research period, research content, and research method related to Christian picture books. For this study, 45 domestic master's and doctoral dissertations were extracted through the National Assembly Library and the Academic Research Information Service (RISS) with the keywords of 'Christian picture book', 'Bible picture book', 'Christian story', and 'Bible story'. The frequency and percentage were calculated by analyzing Christian picture book-related studies according to four criteria: research period, research content, research method, and research subject. As a result of the study, first, the trend of Christian picture book research papers by research period from 1999 to 2021 was 43 master's articles (95.6%) and 2 doctoral articles (4.4%), focusing on Christian picture book-related studies. Second, the trend by research content was found to be 12 basic studies (26.6%) and 33 practical studies (73.4%). Research related to Christian picture books is being actively conducted focusing on practical research rather than basic research. Third, the trend by research method was in the order of 33 quantitative studies (73.4%), 11 literature studies (24.4%), and 1 qualitative study (2.2%). Research related to Christian picture books is centered on quantitative research, and literature research and qualitative research account for a relatively low proportion. Fourth, as for the trends by study subject, there were 35 human subjects (77.8%) and 10 physical subjects (22.2%). Among human subjects, 33 single subjects (73.4%) and 2 mixed subjects (4.4%) were found, and among single subjects, 30 studies (66.7%) targeting children were high. In other words, research on Christian picture books had a higher proportion of studies with children as a single subject than mixed subjects between children and children, children and teachers, and between children and parents.

• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.251-273
• /
• 2022

Recently, research on applying text analysis to deep learning has steadily continued. In particular, researches have been actively conducted to understand the meaning of words and perform tasks such as summarization and sentiment classification through a pre-trained language model that learns large datasets. However, existing pre-trained language models show limitations in that they do not understand specific domains well. Therefore, in recent years, the flow of research has shifted toward creating a language model specialized for a particular domain. Domain-specific pre-trained language models allow the model to understand the knowledge of a particular domain better and reveal performance improvements on various tasks in the field. However, domain-specific further pre-training is expensive to acquire corpus data of the target domain. Furthermore, many cases have reported that performance improvement after further pre-training is insignificant in some domains. As such, it is difficult to decide to develop a domain-specific pre-trained language model, while it is not clear whether the performance will be improved dramatically. In this paper, we present a way to proactively check the expected performance improvement by further pre-training in a domain before actually performing further pre-training. Specifically, after selecting three domains, we measured the increase in classification accuracy through further pre-training in each domain. We also developed and presented new indicators to estimate the specificity of the domain based on the normalized frequency of the keywords used in each domain. Finally, we conducted classification using a pre-trained language model and a domain-specific pre-trained language model of three domains. As a result, we confirmed that the higher the domain specificity index, the higher the performance improvement through further pre-training.