• International Journal of Concrete Structures and Materials
• /
• v.6 no.3
• /
• pp.177-186
• /
• 2012

The temperature distributions of concrete structures strongly depend on the value of thermal conductivity of concrete. However, the thermal conductivity of concrete varies according to the composition of the constituents and the temperature and moisture conditions of concrete, which cause difficulty in accurately predicting the thermal conductivity value in concrete. For this reason, in this study, back-propagation neural network models on the basis of experimental values carried out by previous researchers have been utilized to effectively account for the influence of these variables. The neural networks were trained by 124 data sets with eleven parameters: nine concrete composition parameters (the ratio of water-cement, the percentage of fine and coarse aggregate, and the unit weight of water, cement, fine aggregate, coarse aggregate, fly ash and silica fume) and two concrete state parameters (the temperature and water content of concrete). Finally, the trained neural network models were evaluated by applying to other 28 measured values not included in the training of the neural networks. The result indicated that the proposed method using a back-propagation neural algorithm was effective at predicting the thermal conductivity of concrete.

• Structural Monitoring and Maintenance
• /
• v.5 no.1
• /
• pp.51-65
• /
• 2018

A number of sensing techniques have been implemented for detecting defects in civil infrastructures instead of onsite human inspections in structural health monitoring. However, the issue of faults in sensors has not received much attention. This issue may lead to incorrect interpretation of data and false alarms. To overcome these challenges, this article presents a deep learning-based method with a new architecture of Stateful Long Short Term Memory Neural Networks (S-LSTM NN) for detecting sensor fault without going into details of the fault features. As LSTMs are capable of learning data features automatically, and the proposed method works without an accurate mathematical model. The detection of four types of sensor faults are studied in this paper. Non-stationary acceleration responses of a three-span continuous bridge when under operational conditions are studied. A deep network model is applied to the measured bridge data with estimation to detect the sensor fault. Another set of sensor output data is used to supervise the network parameters and backpropagation algorithm to fine tune the parameters to establish a deep self-coding network model. The response residuals between the true value and the predicted value of the deep S-LSTM network was statistically analyzed to determine the fault threshold of sensor. Experimental study with a cable-stayed bridge further indicated that the proposed method is robust in the detection of the sensor fault.

• Journal of Environmental Science International
• /
• v.21 no.5
• /
• pp.545-553
• /
• 2012

Eco-efficiency considers both environmental impacts and economic values. It is a useful tool for communicating with stakeholders for business decision making. This study evaluated the eco-efficiency factor (EEF) for the energy network of a dyeing company that supplies surplus heat to a neighboring apartment during the night. This symbiosis network is one of the eco-industrial park (EIP) projects in Korea and aims to benefit local residents and the industrial complex by utilizing surplus heat. In this study, two categories were annualized. The first quantified environmental burden based on $CO_2$ emissions and quantified product value in terms of steam sales. The second used a variety of environmental factors, such as fossil fuel, water and waste, to quantify environmental burden and used steam sales to quantify value. The EEF of the symbiosis network was 1.6, using the global warming impact, and determined using the multiple variable, was 1.33. This study shows that the EEF depends on variable details of environmental burden but the values of this project were very high contrast to other business or EIP project.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.12 no.4
• /
• pp.163-169
• /
• 2016

This study is to analyze factors affecting a growth of online community with perspectives of social network. Particularly this study tries to explore structural phenomenon built by interactions between users in a certain of free board belonging to the online community, in which focuses on one's writing a comments responding on those of others. With using SNA(Social Network Analysis), the social network data calculated from users' interaction shown as their comments were collected to draw out each individual's centrality value representing the structure of the online community and also we estimated duration time and the number of each comments as a proxy variable representing growth of the online community. And then cause-and-effect relationship between individual's centrality value and the duration time and the number of each comment were analyzed. As a result of the analysis, Core-Periphery, Centralization and Reciprocity have significant effects on the duration time and the number of each comment, therefore those significant values representing online structure will give an implication to manage, to promote the online community, to forecast its evolution path and to build critical policies.

• The Transactions of the Korea Information Processing Society
• /
• v.2 no.3
• /
• pp.417-424
• /
• 1995

This paper describes MFT(Mean Field Theory) neural network with continuous with continuous variables is applied to quantification analysis problem. A quantification analysis problem, one of the important problems in statistics, is NP complete and arises in the optimal location of objects in the design space according to the given similarities only. This paper presents a MFT neural network with continuous variables for the quantification problem. Starting with reformulation of the quantification problem to the penalty problem, this paper propose a "one-variable stochastic simulated annealing(one-variable SSA)" based on the mean field approximation. This makes it possible to evaluate of the spin average faster than real value calculating in the MFT neural network with continuous variables. Consequently, some experimental results show the feasibility of this approach to overcome the difficulties to evaluate the spin average value expressed by the integral in such models.ch models.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.11 no.4
• /
• pp.227-236
• /
• 1986

In this paper, an integrated voice/data packet network with window flow control is modeled by a colsed multichain queueing system, and its performance is analyzed by the mean value analysis method. Particularly, for the analysis of a packet network having various kinds of messages with different priority classes, we introduce an approach based on the mean value analysis and the concept of effective capacity. By the mathematical analysis and computer simulation, we obtain the following network statistics in the steady state: Mean buffer occupancy at each node, utilization of link throughput of a virtual channel, and mean delay time of each message. Our iterative analysis method can predict the link data status in most cases within about 10 percent of accurady, and the statistics of voice messages and external data within 5 percent as compared to simulation results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.9
• /
• pp.2882-2903
• /
• 2022

Image super-resolution (SR) processing is of great value in the fields of digital image processing, intelligent security, film and television production and so on. This paper proposed a densely connected deep learning network based on cascade architecture, which can be used to solve the problem of super-resolution in the field of image quality enhancement. We proposed a more efficient residual scaling dense block (RSDB) and the multi-channel cascade architecture to realize more efficient feature reuse. Also we proposed a hybrid loss function based on L₁ error and L_∞ error to achieve better L_∞ error performance. The experimental results show that the overall performance of the network is effectively improved on cascade architecture and residual scaling. Compared with the residual dense net (RDN), the PSNR / SSIM of the new method is improved by 2.24% / 1.44% respectively, and the L_∞ performance is improved by 3.64%. It shows that the cascade connection and residual scaling method can effectively realize feature reuse, improving the residual convergence speed and learning efficiency of our network. The L_∞ performance is improved by 11.09% with only a minimal loses of 1.14% / 0.60% on PSNR / SSIM performance after adopting the new loss function. That is to say, the L_∞ performance can be improved greatly on the new loss function with a minor loss of PSNR / SSIM performance, which is of great value in L_∞ error sensitive tasks.

• Korea Trade Review
• /
• v.47 no.4
• /
• pp.137-159
• /
• 2022

This study aims to analyze international trade papers published in Korea during the past 2002-2022 years. Through this study, it is possible to understand the main subject and direction of research in Korea's international trade field. As the research mythologies, this study uses the big data analysis such as the text mining and Social Network Analysis such as frequency analysis, several centrality analysis, and topic analysis. After analyzing the empirical results, the frequency of key word is very high in trade, export, tariff, market, industry, and the performance of firm. However, there has been a tendency to include logistics, e-business, value and chain, and innovation over the time. The degree and closeness centrality analyses also show that the higher frequency key words also have been higher in the degree and closeness centrality. In contrast, the order of eigenvector centrality seems to be different from those of the degree and closeness centrality. The ego network shows the density of business, sale, exchange, and integration appears to be high in order unlike the frequency analysis. The topic analysis shows that the export, trade, tariff, logstics, innovation, industry, value, and chain seem to have high the probabilities of included in several topics.

• The KIPS Transactions:PartD
• /
• v.16D no.6
• /
• pp.851-858
• /
• 2009

Mobbing is not restricted to problem of young people but the bigger recent problem occurs in workspaces. According to reports of ILO and domestic case mobbing in the workplace is increasing more and more numerically from 9.1%('03) to 30.7%('08). These mobbing brings personal and social losses. The proposed algorithm makes it possible to grasp not only current mobbing victims but also potential mobbing victims through user profile and contribute to efficient personnel management. This paper extracts user profile related to mobbing, in a way of selecting seven factors and fifty attributes that are related to this matter. Next, expressing extracting factors as '1' if they are related me or not '0'. And apply similarity function to attributes summation included in factors to calculate similarity between the users. Third, calculate optimizing weight choosing factors included attributes by applying neural network algorithm of SPSS Clementine and through this summation Mobbing-Value(MV) can be calculated . Finally by mapping MV of online social network users to G2 mobbing propensity classification model(4 Groups; Ideal Group of the online social network, Bullies, Aggressive victims, Victims) which is designed in this paper, can grasp mobbing propensity of users, which will contribute to efficient personnel management.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.9
• /
• pp.33-38
• /
• 2014

The FCM based hybrid RBF network is a heterogeneous learning network model that applies FCM algorithm between input and middle layer and applies Max_Min algorithm between middle layer and output. The Max-Min neural network uses winner nodes of the middle layer as input but shows inefficient learning in performance when the input vector consists of too many patterns. To overcome this problem, we propose a dynamic learning rate control based on fuzzy logic. The proposed method first classifies accurate/inaccurate class with respect to the difference between target value and output value with threshold and then fuzzy membership function and fuzzy decision logic is designed to control the learning rate dynamically. We apply this proposed RBF network to the character recognition problem and the efficacy of the proposed method is verified in the experiment.