• Journal of the Korean Society of Marine Environment & Safety
• /
• v.28 no.2
• /
• pp.201-211
• /
• 2022

At the center of the Noryang waterway, the Gwangyang bay area (including the Yeosu Strait) is located at the west, and the Jinju bay area (including Gangjin bay and Sacheon bay) is located at the east. Freshwater from several rivers is flowing into the study area. In particula,r the event of flood, great quantities freshwater flow from Seomjingang (Seomjin river) into the Gwangyang bay area and from Gahwacheon (discharge from Namgang Dam) into the Jinju bay. The Gwangyang and Jinju bay are connected to the Noryang waterway. In addition, freshwater from Seomjingang and Gahwacheon also affect through the Noryang waterway. In this study, we elucidated the characteristics of the tidal exchange rate and residence time for dry season and flood season on 50 frequency, considering freshwater from 51 rivers, including Seomjingang and Gahwacheon, using a particle tracking method. We conducted additional experiments to determine the effect of freshwater from Seomjingang and Gahwacheon during flooding. In both the dry season and flood season, the result showed that the particles released from the Gwangyang bay moved to the Jinju bay through the Noryang waterway. However, comparatively small amount of particles moved from the Jinju bay to the Gwangyang bay. Each experimental case, the sea exchange rate was 44.40~67.21% in the Gwangyang bay and 50.37~73.10% in the Jinju bay, and the average residence time was 7.07~15.36days in the Gwangyang bay and 6.45~12.75days in the Jinju bay. Consequently the sea exchange rate increased and the residence time decreased during flooding. A calculation of cross-section water flux over 30 days for 7 internal and 5 external areas, indicated that the main essential flow direction of the water flux was the river outflow water from Seomjingang flow through the Yeosu strait to the outer sea and from Gahwacheon flow through Sacheon bay, Jinju bay and the Daebang waterway to the outer sea.

• Journal of Practical Agriculture & Fisheries Research
• /
• v.25 no.3
• /
• pp.19-29
• /
• 2023

This study investigated effect of water temperature change on the formation and release of conchospores of Neopyropia yezoensis. We observed that conchocelis growth and conchospores formation in oyster shell at labolatory during 7 weeks. In order to investigate the amount of conchospore formation in oyster shells, which was being cultured at 28℃, was moved to 10℃, 18℃, 28℃, and culture during 6 weeks. At 10℃, we observed an average of 127 for 1 week, 127 for 2 weeks, 95 for 3 weeks, 90 for 4 weeks, 76 for 5 weeks, and 75 for 6 weeks. At 18℃, we observed an average of 141 for 1 week, 135 for 2 weeks, 94 for 3 weeks, 153 for 4 weeks, 162 for 5 weeks, and 2 for 6 weeks. At 28℃, we observed an average of 167 for 1 week, 102 for 2 weeks, 148 for 3 weeks, 157 for 4 weeks, 270 for 5 weeks, and 138 for 6 weeks. Conchospores released from the shell grew into a young thalli in the culture for 6 weeks, and the number of ones was counted. The number of young thalli were investigated at 10℃, 0 for 1 week, 189 for 2 weeks, 200 for 3 weeks, 89 for 4 weeks, 56 for 5 weeks and 27 for 6 weeks. At 18℃, It observed 0 for 1 week, 26 for 2 weeks, 546 for 3 weeks, 16 for 4 weeks, 17 for 5 weeks and 154 for 6 weeks. It was not observed at 28℃.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.18 no.4
• /
• pp.207-222
• /
• 2023

In order to improve employment rates, most universities operate programs to strengthen students' employment and entrepreneurship, regardless of whether they are selected as the Leading Industry-Innovative University (LINC) or not. In particular, in the case of non-metropolitan universities are risking their lives to improve employment rates. In order to overcome the limitations of university establishment type and university location, which absolutely affect the employment rate, we are operating a startup education & startup support program in order to strengthen employment and entrepreneurship, and capstone design & field training as industry-academia-linked education programs are always available. Although there are studies on effectiveness verification centered on LINC (Leaders in Industry-University Cooperation) in previous studies, but a longitudinal study was conducted on all factors of university factors, startup education & startup support, and capstone design & field training as industry-university-linked education programs as factors affecting the employment rate based on public disclosure indicators. No cases of longitudinal studies were reported. This study targets 116 universities that satisfy the conditions based on university disclosure indicators from 2018 to 2020 that were recently released on university factors, startup education & startup support, and capstone design & field training as industry-academia-linked education programs as factors affecting the employment rate. We analyzed the differences between the LINC (Leaders in Industry-University Cooperation) 51 participating universities and 64 non-participating universities. In addition, considering that there is no historical information on the overlapping participation of participating students due to the limitations of public indicators, the Exposure Effect theory states that long-term exposure to employment and entrepreneurship competency enhancement programs will affect the employment rate through competency enhancement. Based on this, the effectiveness of the 2nd LINC+ (socially customized Leaders in Industry-University Cooperation) was verified from 2017 to 2021 through a longitudinal causal relationship analysis. As a result of the study, it was found that the startup education & startup support and capstone design & field training as industry-academia-linked education programs of the 2nd LINC+ (socially customized Leaders in Industry-University Cooperation) did not affect the employment rate. As a result of the longitudinal causal relationship analysis, it was reconfirmed that universities in metropolitan areas still have higher employment rates than universities in non-metropolitan areas due to existing university factors, and that private universities have higher employment rates than national universities. Among employment and entrepreneurship competency strengthening programs, the number of people who complete entrepreneurship courses, the number of people who complete capstone design, the amount of capstone design payment, and the number of dedicated faculty members partially affect the employment rate by year, while field training has no effect at all by year. It was confirmed that long-term exposure to the entrepreneurship capacity building program did not affect the employment rate. Therefore, it was reconfirmed that in order to improve the employment rate of universities, the limitations of non-metropolitan areas and national and public universities must be overcome. To overcome this, as a program to strengthen employment and entrepreneurship capabilities, it is important to strengthen entrepreneurship through participation in entrepreneurship lectures and actively introduce and be confident in the capstone design program that strengthens the concept of PBL (Problem Based Learning), and the field training program improves the employment rate. In order for actually field training affect of the employment rate, it is necessary to proceed with a substantial program through reorganization of the overall academic system and organization.