• Journal of Biomedical Engineering Research
• /
• v.25 no.6
• /
• pp.581-587
• /
• 2004

Conventional power spectrum methods based on FFT, AR method are not appropriate for analyzing biomedical signals whose spectral characteristics change rapidly. On the other hand, time-frequency analysis has more desirable characteristics of a time-varying spectrum. In this study, we investigated the spectral components of heart rate variability(HRV) in time-frequency domain using time frequency analysis methods. In the various time-frequency kernels functions, we studied the suitable kernels for the analysis of HRV using synthetic HRV signals. First, we evaluated the time/frequency resolution and cross term reduction of various kernel functions. Then, from the instantaneous frequency, obtained from time-frequency distribution, the method extracting frequency components of HRV was proposed. Subjects were 17 healthy young men. A coin-stacking task was used to induce mental stress. For each subjects, the experiment time was 3 minutes. Electrocardiogram, measured during the experiment, was analyzed after converted to HRV signal. In the results, emotional stress of subjects produced an increase in sympathetic activity. Sympathetic activation was responsible for the significant increase in the LF/HF ratio. Subjects were divided into two groups with task ability. Subjects who have higher mental stress have lack of task ability.

• Clean Technology
• /
• v.28 no.3
• /
• pp.232-237
• /
• 2022

Biorefineries, in which renewable resources are utilized, are an eco-friendly alternative based on biomass feedstocks. Alginic acid, a major component of brown algae, which is a type of marine biomass, is widely used in various industries and can be converted into value-added chemicals such as sugars, sugar alcohols, furans, and organic acids via catalytic hydrothermal decomposition under certain conditions. In this study, ruthenium-supported activated carbon and magnesium oxide were mixed and applied to the depolymerization of alginic acid in a batch reactor. The addition of magnesium oxide as a basic promoter had a strong influence on product distribution. In this heterogeneous catalytic system, the separation and purification processes are also simplified. After the reaction, low molecular weight alcohols and organic acids with 5 or fewer carbons were produced. Specifically, under the optimal reaction conditions of 30 mL of 1 wt% alginic acid aqueous solution, 100 mg of ruthenium-supported activated carbon, 100 mg of magnesium oxide, 210 ℃ of reaction temperature, and 1 h of reaction time, total carbon yields of 29.8% for alcohols and 43.8% for a liquid product were obtained. Hence, it is suggested that this catalytic system results in the enhanced hydrogenolysis of alginic acid to value-added chemicals.

• Korean Journal of Remote Sensing
• /
• v.39 no.6_3
• /
• pp.1679-1692
• /
• 2023

South Korea is a country that emits a large amount of pollutants as a result of population growth and industrial development and is also severely affected by transboundary air pollution due to its geographical location. As pollutants from both domestic and foreign sources contribute to air pollution in Korea, the location of air pollutant emission sources is crucial for understanding the movement and distribution of pollutants in the atmosphere and establishing national-level air pollution management and response strategies. Based on this background, this study aims to effectively acquire spatial information on domestic and international air pollutant emission sources, which is essential for analyzing air pollution status, by utilizing high-resolution optical satellite images and deep learning-based image segmentation models. In particular, industrial parks and quarries, which have been evaluated as contributing significantly to transboundary air pollution, were selected as the main research subjects, and images of these areas from multi-purpose satellites 3 and 3A were collected, preprocessed, and converted into input and label data for model training. As a result of training the U-Net model using this data, the overall accuracy of 0.8484 and mean Intersection over Union (mIoU) of 0.6490 were achieved, and the predicted maps showed significant results in extracting object boundaries more accurately than the label data created by course annotations.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.139-156
• /
• 2021

The biggest reason for using a deep learning model in image classification is that it is possible to consider the relationship between each region by extracting each region's features from the overall information of the image. However, the CNN model may not be suitable for emotional image data without the image's regional features. To solve the difficulty of classifying emotion images, many researchers each year propose a CNN-based architecture suitable for emotion images. Studies on the relationship between color and human emotion were also conducted, and results were derived that different emotions are induced according to color. In studies using deep learning, there have been studies that apply color information to image subtraction classification. The case where the image's color information is additionally used than the case where the classification model is trained with only the image improves the accuracy of classifying image emotions. This study proposes two ways to increase the accuracy by incorporating the result value after the model classifies an image's emotion. Both methods improve accuracy by modifying the result value based on statistics using the color of the picture. When performing the test by finding the two-color combinations most distributed for all training data, the two-color combinations most distributed for each test data image were found. The result values were corrected according to the color combination distribution. This method weights the result value obtained after the model classifies an image's emotion by creating an expression based on the log function and the exponential function. Emotion6, classified into six emotions, and Artphoto classified into eight categories were used for the image data. Densenet169, Mnasnet, Resnet101, Resnet152, and Vgg19 architectures were used for the CNN model, and the performance evaluation was compared before and after applying the two-stage learning to the CNN model. Inspired by color psychology, which deals with the relationship between colors and emotions, when creating a model that classifies an image's sentiment, we studied how to improve accuracy by modifying the result values based on color. Sixteen colors were used: red, orange, yellow, green, blue, indigo, purple, turquoise, pink, magenta, brown, gray, silver, gold, white, and black. It has meaning. Using Scikit-learn's Clustering, the seven colors that are primarily distributed in the image are checked. Then, the RGB coordinate values of the colors from the image are compared with the RGB coordinate values of the 16 colors presented in the above data. That is, it was converted to the closest color. Suppose three or more color combinations are selected. In that case, too many color combinations occur, resulting in a problem in which the distribution is scattered, so a situation fewer influences the result value. Therefore, to solve this problem, two-color combinations were found and weighted to the model. Before training, the most distributed color combinations were found for all training data images. The distribution of color combinations for each class was stored in a Python dictionary format to be used during testing. During the test, the two-color combinations that are most distributed for each test data image are found. After that, we checked how the color combinations were distributed in the training data and corrected the result. We devised several equations to weight the result value from the model based on the extracted color as described above. The data set was randomly divided by 80:20, and the model was verified using 20% of the data as a test set. After splitting the remaining 80% of the data into five divisions to perform 5-fold cross-validation, the model was trained five times using different verification datasets. Finally, the performance was checked using the test dataset that was previously separated. Adam was used as the activation function, and the learning rate was set to 0.01. The training was performed as much as 20 epochs, and if the validation loss value did not decrease during five epochs of learning, the experiment was stopped. Early tapping was set to load the model with the best validation loss value. The classification accuracy was better when the extracted information using color properties was used together than the case using only the CNN architecture.

• Korean Journal of Environment and Ecology
• /
• v.28 no.6
• /
• pp.627-633
• /
• 2014

Annual biomass production and amount of organic carbon in Abies koreana forest at Mt. Halla were conducted as a part of Korea National Long-Term Ecological Research (KNLTER). We measured standing biomass change of litter, soil production and organic carbon amounts of the forest floor and soil layer of A. koreana forest in Mt. Halla from 2009 to 2013 in permanent plots. Standing biomass, which was determined by allometric method, was converted into $CO_2$. The standing biomass in A. koreana forest was 98.88, 106.42, 107.67, 108.31, $91.48ton\;ha^{-1}$ in 2009, 2010, 2011, 2012 and 2013 year, respectively. The amount of annual carbon allocated to above ground was 35.95, 38.69, 38.96, 39.46, $33.2ton\;C\;ha^{-1}$ and below ground biomass was 8.54, 9.2, 9.49, 9.28, $7.97ton\;C\;ha^{-1}$ in 2009, 2010, 2011, 2012 and 2013 year, respectively. Amount of organic carbon returned to the forest via litterfall was 1.09, 1.80, 1.32, 2.46 and $1.20ton\;C\;ha^{-1}$ in 2009, 2010, 2011, 2012 and 2013. Amount of organic carbon in annual litter layer on forest floor was 2.74, 2.43, 2.00 and $1.16ton\;C\;ha^{-1}$ in 2010, 2011, 2012 and 2013 year, respectively. Amount of organic carbon within 20cm soil depth was 55.77, 54.9, 50.69, 44.42 and $41.87ton\;C\;ha^{-1}20cm^{-1}$ in 2009, 2010, 2011, 2012 and 2013 year, respectively. Then standing biomass and organic carbon distribution increased steadily until 2012. But there declined in 2013 because of the typhoon Bolaven. Thus, standing biomass and organic carbon distribution of this subalpine forest were largely affected by natural disturbance factor.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.167-181
• /
• 2018

Over the past decade, deep learning has been in spotlight among various machine learning algorithms. In particular, CNN(Convolutional Neural Network), which is known as the effective solution for recognizing and classifying images or voices, has been popularly applied to classification and prediction problems. In this study, we investigate the way to apply CNN in business problem solving. Specifically, this study propose to apply CNN to stock market prediction, one of the most challenging tasks in the machine learning research. As mentioned, CNN has strength in interpreting images. Thus, the model proposed in this study adopts CNN as the binary classifier that predicts stock market direction (upward or downward) by using time series graphs as its inputs. That is, our proposal is to build a machine learning algorithm that mimics an experts called 'technical analysts' who examine the graph of past price movement, and predict future financial price movements. Our proposed model named 'CNN-FG(Convolutional Neural Network using Fluctuation Graph)' consists of five steps. In the first step, it divides the dataset into the intervals of 5 days. And then, it creates time series graphs for the divided dataset in step 2. The size of the image in which the graph is drawn is $40(pixels){\times}40(pixels)$, and the graph of each independent variable was drawn using different colors. In step 3, the model converts the images into the matrices. Each image is converted into the combination of three matrices in order to express the value of the color using R(red), G(green), and B(blue) scale. In the next step, it splits the dataset of the graph images into training and validation datasets. We used 80% of the total dataset as the training dataset, and the remaining 20% as the validation dataset. And then, CNN classifiers are trained using the images of training dataset in the final step. Regarding the parameters of CNN-FG, we adopted two convolution filters ($5{\times}5{\times}6$ and $5{\times}5{\times}9$) in the convolution layer. In the pooling layer, $2{\times}2$ max pooling filter was used. The numbers of the nodes in two hidden layers were set to, respectively, 900 and 32, and the number of the nodes in the output layer was set to 2(one is for the prediction of upward trend, and the other one is for downward trend). Activation functions for the convolution layer and the hidden layer were set to ReLU(Rectified Linear Unit), and one for the output layer set to Softmax function. To validate our model - CNN-FG, we applied it to the prediction of KOSPI200 for 2,026 days in eight years (from 2009 to 2016). To match the proportions of the two groups in the independent variable (i.e. tomorrow's stock market movement), we selected 1,950 samples by applying random sampling. Finally, we built the training dataset using 80% of the total dataset (1,560 samples), and the validation dataset using 20% (390 samples). The dependent variables of the experimental dataset included twelve technical indicators popularly been used in the previous studies. They include Stochastic %K, Stochastic %D, Momentum, ROC(rate of change), LW %R(Larry William's %R), A/D oscillator(accumulation/distribution oscillator), OSCP(price oscillator), CCI(commodity channel index), and so on. To confirm the superiority of CNN-FG, we compared its prediction accuracy with the ones of other classification models. Experimental results showed that CNN-FG outperforms LOGIT(logistic regression), ANN(artificial neural network), and SVM(support vector machine) with the statistical significance. These empirical results imply that converting time series business data into graphs and building CNN-based classification models using these graphs can be effective from the perspective of prediction accuracy. Thus, this paper sheds a light on how to apply deep learning techniques to the domain of business problem solving.

• Journal of Korean Society of Transportation
• /
• v.23 no.1
• /
• pp.93-102
• /
• 2005

An Origin-Destination (O/D) is considered as one of the important information in route choices and trip assignments. A household interview survey is deemed to be the traditional and the most widely used method in making sample O/D and its conversion to the total O/D. Some researchers have studied to estimate dynamic O/D from the relationship between link volumes and trip assignment model. Nowadays, owing to the recent rapid spread of cellular phones. Location information of the cellular phone through the Base Station(BS) is considered as an alternative to O/D estimation. In this study, the methodology of generating BS-based O/D and the methodology of converting this O/D into an administrative district-based O/D are proposed. The information of GPS positions and cellular BS positions have acquired by establishing GPS equipment and cellular phone on taxies in Cheongju. Three weeks data are collected and used in estimating O/D by matching them on a digital map. Scatter diagram and sample correlation coefficients are used to investigate the similarity of the GPS-based O/D pattern among weeks, among days, and among times in day. The results show that there are few significant differences among weeks. But there is a difference in O/C pattern between weekday and weekend. Furthermore, there is a difference between morning peak and afternoon peak. Two methodologies are proposed to convert BS-based O/D into an administrative district-based O/D. The first one is to use the distribution pattern of GPS coordinates, the other is to use the coverage area of the BSs. To validate such converted O/D, GPS O/D is used as a true value. The statical analyses through scatter diagram, MAE and RMSE shows that there is few significant defference of pattern between the estimated BS-based O/D and GPS O/D. In the case of using only cellular information, the methodology using coverage area of the BSs is recommended for estimating O/D.

• The Korean Journal of Archival Studies
• /
• no.5
• /
• pp.39-85
• /
• 2002

This paper is aimed at proposing the policies for managing archives in the process of promoting Electronic Government System. Although there have been many studies of electronic government project and plans for its establishment, this research examines the electronic government system and its problems on the basis of archival science. What I acquired in this paper is as follows. The development of information technology needs great changes ranging from the nation to the individuals. It becomes common that the use of computerized program for business purposes, computerization of information materials and the effective way of search use of electronic documents. Therefore, more and more countries all over the world have been seeking to promote 'Electronic Government', which applies the fruits of the development in information technology to administration process. Recently, Korea has been rapidly entered into the 'Electronic Government' system being against the traditional way of administration. In electronic government system, the 'Life Cycle' of public records will be computerized. Therefore, it is important to change and develop along with the government's policies for 'electronic government project' in the archival management system. This means that the archival management system which have put emphasis on the textual records should be converted to electronic records system. In other words, the records management in electronic government system requires not the transfer and preservation of the records but the consistent management system including the whole process of creating, appraising, arranging, preserving and using the records. So, the systematic management of electronic records plays an important role in realization of electronic government, but it is a subject to be realized by electronic government at the same time. However, the government have overlooked the importance of archival management for long time, especially the importance of electronic records management system. First of all, this research attempts to infer limits and problems through the theoretical considerations of the existing studies for electronic government and to clear up the relations between electronic government and archival management. Based on this, I'll seek to progress the study through reviewing the present condition of archival management in the process of promoting electronic government and suggesting the policies for enhancing the successful electronic government and the construction of scientific archival management system. Since early 1990, many countries in the world have been making every effort to concrete 'Electronic Government'. Using the examples in other nations, it is not difficult to recognize that the embodiment of electronic government is closely connected with the archival management policies. Korea have completed legal and institutional equipments including the new establishment of "Electronic Government Law" to realize electronic government. Also, Korea has been promoting electronic government with the Ministry of Government Administration and Home Affairs and Government Computer Center as a leaders. Though managing records, especially the management of electronic records is essential in electronic government system, we haven't yet discussed this section in Korea. This is disapproved by the fact the Government Archives and Records Service has played little role in promoting electronic government project. There are two problems relating this environment. First, present system can't meet the consistent 'Life Cycle' ranging from the creation to the preservation of electronic records. Second, the 'Life Cycle' of electronic records is divided into two parts and managed separately by GCC and GARS. The life of records is not end with the process raged from creation to distribution. On the other hand, the records are approved their value only whole procedures. Therefore, GARS should play a deading role in designing and establishing the archival management system. The answer to these problems, is as follows. First, we have to complete the electronic records management system through introducing ERMS not EDMS. This means that we should not change and develop towards ERMS simply with supplementing the current electronic records management system. I confirm that it is important and proper to establish ERMS system from the very beginning of the process of promoting electronic government. Second, I suggest the developmental integration of GARS and GCC. At present, the divided operations of GCC and GARS, the former is in charge of the management center for electronic business and the latter is the hub institution of managing nation's records and archives result in many obstacles in establishing electronic government system and accomplishing the duties of systematic archival management. Therefore, I conclude that the expansive movement towards 'National Archives' through the integration among the related agencies will make a great contribution to the realization of electronic government and the establishment of archival management system. In addition to this, it will be of much help to constitute and operate the 'Task Force' regarding the management of electronic records with the two institution as the central figures.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.23 no.2
• /
• pp.8-8
• /
• 1987

This paper describes on the distribution of the vibration and the noise produced on a skipjack pole and line training ship M/S Pusan 403 (243GT, 1,000ps) under the cruising or drifting condition. The vibration and the noise level were measured by use of protable vibration analyzer (B and K 3513) and sound level meter (B and K 2205), and so the vibration level was converted into dB unit. The check points were set through every decks and around important places of the ship. The results obtained can be summarized as follows: 1. The vibration and the noise level 1) On the main deck, both the vibration and the noise level were highest at the vertically above the main engine, whereas the vibration level was the lowest in the bow store and the noise level beneath the bridge. 2) Under cruising condition, the vibration level around the cylinder head of main engine, port side of the engine room, on the shaft tunnel was 80, 67, 65 dB and the noise level 104, 87, 86 dB, respectively. 3) The vibration level on the vertical line passing through the bridge was the highest at the orlop deck with 60 dB and the lowest on the bridge deck with 55 dB, whereas the noise level the highest at the compass deck with 75 dB and the lowest at the orlop deck with 53 dB. 4) The vibration and the noise level on the open decks were the highest with 65 dB and 84 dB on the boat deck, whereas the vibration level was the lowest at the lecture room with 51 dB and the noise level the lowest at the fore castle deck with 57 dB. 5) On the orlop decks, both the vibration and the noise level were the highest at the engine room with 65 dB and 85 dB, and the lowest at bow store with 54 dB and 52 dB, respectively. Comparing with the vibration level and the noise level, the vibration level was higher than the noise level in the bow part and it was contrary in the stern part of the ship. 2. Vibration analysis 1) The vibration displacement and the vibration velocity were the greatest at the cylinder head of main engine with 100μm and 11mm/sec, and were the smallest at the compass deck with 3μm and 0.07mm/sec. They were also attenuated rapidly around the frequency of 100Hz and over. 2) The vibration acceleration was the greatest at the cylinder head with the main frequency of 1KHz and the acceleration of 1.1mm/sec super(2), and the smallest at the compass deck with 30KHz and 0.05mm/sec super(2).

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.23 no.2
• /
• pp.54-60
• /
• 1987

This paper describes on the distribution of the vibration and the noise produced on a skipjack pole and line training ship M/S Pusan 403 (243GT, 1,000ps) under the cruising or drifting condition. The vibration and the noise level were measured by use of protable vibration analyzer (B and K 3513) and sound level meter (B and K 2205), and so the vibration level was converted into dB unit. The check points were set through every decks and around important places of the ship. The results obtained can be summarized as follows: 1. The vibration and the noise level 1) On the main deck, both the vibration and the noise level were highest at the vertically above the main engine, whereas the vibration level was the lowest in the bow store and the noise level beneath the bridge. 2) Under cruising condition, the vibration level around the cylinder head of main engine, port side of the engine room, on the shaft tunnel was 80, 67, 65 dB and the noise level 104, 87, 86 dB, respectively. 3) The vibration level on the vertical line passing through the bridge was the highest at the orlop deck with 60 dB and the lowest on the bridge deck with 55 dB, whereas the noise level the highest at the compass deck with 75 dB and the lowest at the orlop deck with 53 dB. 4) The vibration and the noise level on the open decks were the highest with 65 dB and 84 dB on the boat deck, whereas the vibration level was the lowest at the lecture room with 51 dB and the noise level the lowest at the fore castle deck with 57 dB. 5) On the orlop decks, both the vibration and the noise level were the highest at the engine room with 65 dB and 85 dB, and the lowest at bow store with 54 dB and 52 dB, respectively. Comparing with the vibration level and the noise level, the vibration level was higher than the noise level in the bow part and it was contrary in the stern part of the ship. 2. Vibration analysis 1) The vibration displacement and the vibration velocity were the greatest at the cylinder head of main engine with 100$\mu$m and 11mm/sec, and were the smallest at the compass deck with 3$\mu$m and 0.07mm/sec. They were also attenuated rapidly around the frequency of 100Hz and over. 2) The vibration acceleration was the greatest at the cylinder head with the main frequency of 1KHz and the acceleration of 1.1mm/sec super(2), and the smallest at the compass deck with 30KHz and 0.05mm/sec super(2).