• 한국정보교육학회:학술대회논문집
• /
• 2021.08a
• /
• pp.419-424
• /
• 2021

The purpose of this study is to understand the educational context including the actual learning process and learner perception using the embodied metaphor in AI education. To this end, a class was designed to utilize the embodied metaphor-based unplugged activity through a qualitative approach. Matrix analysis technique was used to analyze the data collected throughout the course of the class to analyze the experiences and perceptions according to the characteristics of the learner, and the learning context. The results of the study were: First, there was a difference according to the learner's prior experience in the effect on the representative knowledge and the subsequent practice process. Next, the embodied metaphor-based unplugged activity showed soft landing effects on practice and text coding. Finally, the organic integration of unplugged and plugged-in classes helped learners understand the potential of computational thinking.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.59-59
• /
• 2022

세계적으로 지구온난화를 동반한 기후변화로 인해 자연재난이 빈번하게 발생하고 있다. 재해의 발생 유형 중 집중호우와 태풍으로 인한 수문학적 재해가 대부분을 차지하고 있다. 이와 같이 홍수로 인해 발생하는 피해는 강우의 특성과 지역적 특성에 따라 피해의 규모와 범위가 달라진다. 따라서 이러한 이질적인 홍수피해로부터 재산과 인명을 보호하기 위해서는 위해성(Hazard), 노출성(Exposure), 취약성(Vulnerability)을 고려하여 지역 특성에 맞는 홍수방어계획을 수립해야한다. 본 연구에서는 전국 228개 행정구역을 대상으로 과거에 실제로 발생하였던 홍수피해 사례 조사를 통해 지역별 홍수피해 특성을 파악하여 지역 특성을 고려한 홍수피해 강우기준을 제시하고자 하였다. 이를 위해서 재해연보 보고서에 기재되어 있는 과거 홍수피해 기간과 홍수피해액을 수집하였고, 홍수피해 기간동안의 강우량과 뉴스 기사를 수집하여 뉴스 기사에서 언급되었던 홍수피해 현상 정보를 수집하였다. 수집된 홍수피해 정보를 통해 지역별 노출성과 취약성이 반영된 현상기반 강우등급을 제시하였으며, 이와 함께 지역별 강우특성을 나타내며 위해성을 내포하고 있는 확률강우량과의 합성을 통해 위해성, 노출성, 취약성을 고려한 지역별 홍수피해 강우기준을 제시하였다. 대부분 홍수피해에 관한 정보를 재해연보 보고서를 활용하여 수집하지만 홍수피해 현상에 대한 정보를 포함하고 있지 않기 때문에 지역별로 홍수피해로부터 발생하는 홍수피해 유형에 대해 파악하기에는 한계가 있다. 따라서 본 연구에서는 과거 홍수피해가 발생했던 기간에 대해 뉴스 기사를 수집하여 홍수피해 현상 정보를 수집하였고, 수집된 홍수피해 현상 정보를 텍스트 마이닝(Text Mining) 기법을 적용하여 홍수피해 현상 키워드 빈도분석을 통해 어떠한 홍수피해 유형에 취약한지 파악하였다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.232-234
• /
• 2022

Big data is widely used in customer relationship management, relationship marketing, financial business improvement, credit information and risk management. Moreover, as non-face-to-face financial transactions have become more active recently due to the COVID-19 virus, the use of financial big data is more demanded in terms of relationships with customers. In terms of customer relationship, financial big data has arrived at a time that requires an emotional rather than a technical approach. In relational marketing, it was necessary to emphasize the emotional aspect rather than the cognitive, rational, and rational aspects. Existing traditional financial data was collected and utilized through text-type customer transaction data, corporate financial information, and questionnaires. In this study, the customer's emotional image data, that is, atypical data based on the customer's cultural and leisure activities, is acquired through SNS and the customer's activity image is analyzed with an artificial intelligence CNN algorithm. Activity analysis is again applied to the annotated AI, and the AI big data model is designed to analyze the behavior model shown in the annotation.

• Annual Conference of KIPS
• /
• 2008.11a
• /
• pp.603-606
• /
• 2008

최근 들어 분자생물학의 급속한 발전과 2001년 인간유전체사업의 완료로 인해 전세계적으로 엄청난 양의 유전정보가 공개되었다. 유전자 서열정보는 그 양이 방대하고 다양하기에 데이터베이스 구축 및 분석을 위하여 고성능 컴퓨터 및 정보기술 기법이 필요하다. 그래서 컴퓨터를 활용하여 생물학적 데이터를 수집, 관리, 저장, 평가, 분석하는 연구분야인 생명정보학(바이오인포매틱스)이라는 학문이 지속적으로 발전하고 있다. 이런 생명정보학 발전에 발맞추어 한국과학기술정보연구원(KISTI)에서는 정보기술을 기반으로 한 생명정보 인프라를 구축하여 생명과학 연구자들에게 제공하고 있다. 본 논문에서는 생명정보 데이터베이스중에서 연구자들이 가장 많이 이용하는 유전자 데이터베이스인 Genbank를 활용 및 분석하여 KISTI에서 운영하는 학술논문 제공 사이트인 NDSL(http://scholar.ndsl.kr)과 연계 가능한 서비스 모델을 개발하기 위하여 1) NCBI FTP 사이트에서 Genbank 데이터를 수집하고, 2) Genbank 텍스트 파일을 유전자 기본정보와 참고 데이터베이스로 재구축하며, 3) Genbank refrence 필드에서 논문 및 특허 정보 추출을 통한 새로운 테이블을 생성하여 NDSL과 연계 가능한 서비스 모델을 제안하였다.

• Asia-Pacific Journal of Business
• /
• v.14 no.3
• /
• pp.241-250
• /
• 2023

Purpose - The purpose of this study was to investigate the perception of 'unmanned cafes' on the network through big data analysis, and to identify the latest trends in rapidly changing consumer perception. Based on this, I would like to suggest that it can be used as basic data for the revitalization of unmanned cafes and differentiated marketing strategies. Design/methodology/approach - This study collected documents containing unmanned cafe keywords for about three years, and the data collected using text mining techniques were analyzed using methods such as keyword frequency analysis, centrality analysis, and keyword network analysis. Findings - First, the top 10 words with a high frequency of appearance were identified in the order of unmanned cafes, unmanned cafes, start-up, operation, coffee, time, coffee machine, franchise, and robot cafes. Second, visualization of the semantic network confirmed that the key keyword "unmanned cafe" was at the center of the keyword cluster. Research implications or Originality - Using big data to collect and analyze keywords with high web visibility, we tried to identify new issues or trends in unmanned cafe recognition, which consists of keywords related to start-ups, mainly deals with topics related to start-ups when unmanned cafes are mentioned on the network.

• Journal of Korea Society of Industrial Information Systems
• /
• v.29 no.3
• /
• pp.79-87
• /
• 2024

Recently, the importance of pre-trained language models (PLM) has been emphasized for natural language processing (NLP) such as text classification, sentiment analysis, and question answering. Korean PLM shows high performance in NLP in general-purpose domains, but is weak in domains such as finance, medicine, and law. The main goal of this study is to propose a language model learning process and method to build a financial-specific language model that shows good performance not only in the financial domain but also in general-purpose domains. The five steps of the financial-specific language model are (1) financial data collection and preprocessing, (2) selection of model architecture such as PLM or foundation model, (3) domain data learning and instruction tuning, (4) model verification and evaluation, and (5) model deployment and utilization. Through this, a method for constructing pre-learning data that takes advantage of the characteristics of the financial domain and an efficient LLM training method, adaptive learning and instruction tuning techniques, were presented.

• Journal of the Korea Institute of Building Construction
• /
• v.24 no.1
• /
• pp.121-131
• /
• 2024

This research explores the potential of leveraging unstructured data from construction supervision documents, which contain detailed inspection insights from independent third-party monitors of building construction processes. With the evolution of analytical methodologies, such unstructured data has been recognized as a valuable source of information, offering diverse insights. The study introduces a framework designed to assess cost performance by applying advanced analytical methods to the unstructured data found in final construction supervision reports. Specifically, key phrases were identified using text mining and social network analysis techniques, and these phrases were then analyzed through binomial logistic regression to assess cost performance. The study found that predictions of cost performance based on unstructured data from supervision documents achieved an accuracy rate of approximately 73%. The findings of this research are anticipated to serve as a foundational resource for analyzing various forms of unstructured data generated within the construction sector in future projects.

• Journal of the Korean Society for information Management
• /
• v.41 no.2
• /
• pp.269-294
• /
• 2024

This study analyzed the news media coverage of QRcodes in Korea over a 16-year period (2008 to 2023). A total of 13,335 articles were extracted from the Korea Press Foundation's BigKinds. A quantitative and content analysis was conducted on the news frames. The results indicated that the quantity of news coverage has increased. The greatest quantity of news coverage was observed in 2020, and the most frequently discussed topic in the news was 'IT_Science'. The results of the keyword analysis indicated that the primary words were 'QRcode', 'smartphone', 'service', 'application', and 'payment'. The news media primarily focused on the QRcode's ability to provide instant access and recognition technology. This study demonstrates that advanced information and communication technologies and the increased prevalence of mobile devices have led to a rise in the utilization of QRcodes. Furthermore, QRcodes have become a significant information media in contemporary society.

• Journal of Navigation and Port Research
• /
• v.48 no.1
• /
• pp.1-6
• /
• 2024

Aid-to-Navigation facility is important for maritime traffic safety. In Korea, for safe maritime traffic, the Ministry of Oceans and Fisheries is using an Aid-to-Navigation management system. The current Aid-to-Navigation management system displays information based on text, making it difficult to determine the impact if Aid-to-Navigation fails or an accident occurs. A simulator can be used to verify the placement of Aid-to-Navigation. However, real-time information is not applied and maintenance of the simulator is expensive. Additionally, the Aid-to-Navigation simulator cannot simulate effects of port backlighting. To improve these issues, we proposed an Aid-to-Navigation management platform based on digital twin technology. This system can predict failures by analyzing real-time sensor data collected from navigation signs. We plan to develop a function that can simulate Aid-to-Navigation placement. Aid-to-Navigation is expected to be managed efficiently by applying digital twin technology.

• Convergence Security Journal
• /
• v.23 no.4
• /
• pp.61-70
• /
• 2023

Clustering is an unsupervised learning method that involves grouping data based on features such as distance metrics, using data without known labels or ground truth values. This method has the advantage of being applicable to various types of data, including images, text, and audio, without the need for labeling. Traditional clustering techniques involve applying dimensionality reduction methods or extracting specific features to perform clustering. However, with the advancement of deep learning models, research on deep clustering techniques using techniques such as autoencoders and generative adversarial networks, which represent input data as latent vectors, has emerged. In this study, we propose a deep clustering technique based on deep learning. In this approach, we use an autoencoder to transform the input data into latent vectors, and then construct a vector space according to the cluster structure and perform k-means clustering. We conducted experiments using the MNIST and Fashion-MNIST datasets in the PyTorch machine learning library as the experimental environment. The model used is a convolutional neural network-based autoencoder model. The experimental results show an accuracy of 89.42% for MNIST and 56.64% for Fashion-MNIST when k is set to 10.