• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.472-475
• /
• 2020

본 논문에서는 청각장애인에게 자막방송을 제공하기 위하여 오디오 콘텐츠에 등장하는 음향 사건을 분류하는 기법을 제안한다. 제안된 기법은 복수의 잔차 신경망(ResNet)을 연결하는 연쇄잔차(concatenated residual) 신경망 구조를 갖는다. 신경망의 입력 특징을 위해 음성의 멜-주파수 켑스트럼 벡터를 다수의 프레임으로 결합하여 형성한 2 차원 이미지와 전체 프레임에 대한 멜-주파수 켑스트럼 벡터들로부터 얻은 1 차원의 통계 특징벡터를 얻는다. 각각의 입력은 2 차원 잔차 신경망과 1 차원 잔차 신경망으로 모델링되고, 두 개의 잔차 신경망을 연쇄연결(concatenation)하는 구조를 가진 연쇄잔차 신경망으로 구성된다. 성능평가를 위해 수집된 데이터셋으로부터 6-fold 교차검증을 통해 평가한 결과, 85.48%의 분류 정확도를 얻을 수 있었다.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.6
• /
• pp.1065-1071
• /
• 2023

In this paper discuss with research developing an LSTM model for classifying the behavior of occupants within a residence. The multi-sensor consists of an IAQ (Indoor Air Quality) sensor that measures indoor air quality, a UWB radar that tracks occupancy detection and location, and a Piezo sensor to measure occupants' biometric information, and collects occupant behavior data such as going out, staying, cooking, cleaning, exercise, and sleep by constructed an experimental environment similar to the actual residential environment. After the data with removed outliers and missing, the LSTM model is used to calculate accuracy, sensitivity, specificity of the occupant behavior classification model, T1 score.

• Journal of Digital Convergence
• /
• v.11 no.4
• /
• pp.91-99
• /
• 2013

While there is the active integration of various industries, a convergence of the software and knowledge service industries including software used in finance and counseling products is creating the necessity to include software industry utilization sectors aside from covering only software products and service production activities. Furthermore, to cope with the radical environment changes in the software industry when it comes to categorizing mobile and cloud computing areas into a software and classification system, we are at a point where there is a need to establish a directional nature on what should be included. In order to establish an integrated classification of newly introduced technologies, products and services, this paper aims to discover areas not included in the classification standard because of the ecological characteristics of the software. It also wants to differentiate the classification system and identify its incomplete areas such as the lack of connections within the system to ultimately establish such for newly surfacing software fields.

• Journal of Information Management
• /
• v.41 no.1
• /
• pp.173-200
• /
• 2010

National construction project for SOC(Social Overhead Capital) is led by the government and invested high cost and all capabilities for long times. In general case the life times of SOC facilities are permanently. According to the long life time of the facilities documents and records of national projects are also retained permanently. Consistent classification systems are required to operation and maintenance of the facilities efficiently and to support the organic co-works for several Stakeholders. Therefore the classification of national construction projects are based on WBS of the project. WBS is the hierarchy structure that established and developed in project management project management methodology to produce deliverables of the projects. As a result, this study provides a prospect of project records classification systems in the 21st Century.

• Journal of Information Management
• /
• v.39 no.4
• /
• pp.47-66
• /
• 2008

This study introduced the machine learning algorithms to overcome the many different limitations involved with manual classification and to provide the users with faster and more accurate classification service. The experiments objects of the study were consisted of 100 literature titles for each of the eight subject categories in MeSH. The algorithms used to the experiments included Neural network, C5.0, CHAID and KNN. As results, the combination of the neural network and C5.0 technique recorded classification accuracy of 83.75%, which was 2.5% and 3.75% higher than that of the neural network alone and C5.0 alone, respectively. The number represented the highest accuracy rates among the four classification experiments. Thus the use of the neural network and C5.0 technique together will result in higher accuracy rates than the techniques individually.

• Journal of Information Management
• /
• v.43 no.1
• /
• pp.41-62
• /
• 2012

Because of the electronic commerce development, share of the internet bookstore is rising every year in the publication circulation market. Also, the book search in the internet bookstore may smarten more forward according to development trend of the electronic book, and importance of the classification system of the internet bookstore is expected that also rise continuously. The purpose of this study is to offer the classification system that can search the social science books of the internet bookstore efficiently. For this study, compared and analyzed the classification system of internal and external 10 internet bookstores. This study may be used to other subject study of the internet bookstore, and it expect to is used to basic data for the classification system of the internet bookstore in the future.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.10
• /
• pp.6379-6387
• /
• 2014

Standardized construction information classification enables extensive information from the construction industry to be shared and exchanged clearly by means of unified information expression. Recently, the constriction information classification in North America and Europe has been revised to adapt to BIM technology and applied to various BIM projects. In Korea, government established construction information classification and made efforts to be standardized. On the other hand, it showed low applicability at the practice field and had no revision plan for adapting BIM. In this study, the authors suggest an improvement plan of construction information classification for applying BIM with the purpose of availability enhancement and adaption to a construction environment change.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2014.10a
• /
• pp.331-337
• /
• 2014

Paradigm shift to cloud computing has increased the importance of security. Even though public cloud computing providers such as Amazon, already provides security related service like firewall and identity management services, it is not suitable to protect data in cloud environments. Because in public cloud computing environments do not allow to use client's own security solution nor equipments. In this environments, user are supposed to do something to enhance security by their hands, so the needs of visualized security management arises. To implement visualized security management, developing near realtime data handling & packet classification mechanisms are crucial. The key technical challenges in packet classification is how to classify packet in the manner of unsupervised way without human interactions. To achieve the goal, this paper presents automated packet classification mechanism based on naive-bayesian and packet Chunking techniques, which can identify signature and does machine learning by itself without human intervention.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.4
• /
• pp.382-390
• /
• 2019

As there is severe competition in the global foldable display market, strategic technology planning is required. Patent analysis as a tool for technology planning has frequently been used due to data characteristics such as openness, formality, and detailed information. However, traditional patent analysis has various limitations such as quantitative approaches are limited in evaluating contents of patents and identifying core technologies of firms as they rely on number of patents, and qualitative approaches have time and cost problems as researchers have to investigate each patent on a case-by-case basis. In this research, we analyze core technologies of firms in the foldable display sector analyzing patent co-classification Network. Results show that the number of patent applications has rapidly increased since 2014, and 92% of these patents are held by two panel manufacturers, SDC and LGD, and two device manufacturers, SEC and LGE. Network analysis shows that the two panel manufacturers' core technologies are similar and two device manufacturers are notably different. This research provides implications to the sector. Moreover, this study provides unique results drawn from co-classification network analysis, and therefore, our research suggests patent co-classification analysis as an effective tool for technology planning.

• The Journal of Korean Association of Computer Education
• /
• v.7 no.4
• /
• pp.51-60
• /
• 2004

This paper proposes an automated coding system of Korean standard industry/occupation for census which reduces a lot of cost and labor for manual coding. The proposed system converts natural language responses on survey questionnaires into corresponding numeric codes using information retrieval techniques and document classification algorithm. The system was experimented with 46,762 industry records and occupation 36,286 records using 10-fold cross -validation evaluation method. As experimental results, the system show 87.08% and 66.08% production rates when classifying industry records into level 2 and level 5 codes respectively. The system shows slightly lower performances on occupation code classification. We expect that the system is enough to be used as a semi-automate coding system which can minimize manual coding task or as a verification tool for manual coding results though it has much room to be improved as an automated coding system.