• Journal of Korea Artificial Intelligence Association
• /
• v.1 no.1
• /
• pp.7-10
• /
• 2023

This study proposes an AI-based shoe recommendation model based on user clothing image data to solve the problem of the global fashion industry, which is worsening due to factors such as the economic downturn. Shoes are an important part of modern fashion, and this research aims to improve user satisfaction and contribute to economic growth through a generative AI-based shoe recommendation service. By utilizing generative AI in the personalized consumer market, we show the feasibility, efficiency, and improvements through an accessible web-based implementation. In conclusion, this study provides insights to help fulfill consumer needs in the ever-changing fashion market by implementing a generative AI-based shoe recommendation model.

• Journal of Platform Technology
• /
• v.11 no.6
• /
• pp.3-12
• /
• 2023

In today's system operation, it is difficult to detect failures and take immediate action in the case of a shortage of manpower compared to the number of equipment or failures in vulnerable time zones, which can lead to delays in failure recovery. In addition, various algorithms exist to detect abnormal symptom data, and it is important to select an appropriate algorithm for each problem. In this paper, an ensemble-based isolation forest model was used to efficiently detect multivariate point anomalies that deviated from the mean distribution in the data set generated to predict system failure and minimize service interruption. And since significant changes in memory space usage are observed together with changes in CPU usage, the problem is solved by using LSTM-Auto Encoder for a collective anomaly in which another feature exhibits an abnormal pattern according to a change in one by comparing two or more features. did In addition, evaluation indicators are set for the performance evaluation of the model presented in this study, and then AI model evaluation is performed.

• The Korean Journal of Applied Statistics
• /
• v.32 no.5
• /
• pp.693-702
• /
• 2019

Deep learning has gained popularity for the classification and prediction task. Neural network layers become deeper as more data becomes available. Saturation is the phenomenon that the gradient of an activation function gets closer to 0 and can happen when the value of weight is too big. Increased importance has been placed on the issue of saturation which limits the ability of weight to learn. To resolve this problem, Glorot and Bengio (Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 249-256, 2010) claimed that efficient neural network training is possible when data flows variously between layers. They argued that variance over the output of each layer and variance over input of each layer are equal. They proposed a method of initialization that the variance of the output of each layer and the variance of the input should be the same. In this paper, we propose a new method of establishing initialization by adopting truncated normal distribution and truncated cauchy distribution. We decide where to truncate the distribution while adapting the initialization method by Glorot and Bengio (2010). Variances are made over output and input equal that are then accomplished by setting variances equal to the variance of truncated distribution. It manipulates the distribution so that the initial values of weights would not grow so large and with values that simultaneously get close to zero. To compare the performance of our proposed method with existing methods, we conducted experiments on MNIST and CIFAR-10 data using DNN and CNN. Our proposed method outperformed existing methods in terms of accuracy.

• Journal of the Korean Society for Marine Environment & Energy
• /
• v.12 no.4
• /
• pp.255-263
• /
• 2009

This study assessed the facility volume distribution for artificial reefs (ARs) deployed at the sea bottom in fishing grounds in the Busan Sea region, and estimated the efficient use of volume in the initial AR group. Analysis of the volume of AR facilities in the Busan Sea region indicated that the quantity and surface areas of deployed facilities varied greatly between 1997 and 2004. However, a rapid decrease in AR facility volume after 2003 indicated an increased use of different kinds of ARs, such as steel or ceramic ARs. And the results revealed that a total of 2,777 AR facilities were deployed in the Gijang sea region with a founding ratio of 89%; corresponding numbers for the Heaundae and Namgu sea region were 7,691 and 84.9%, and 905 and 96% for the cubic type of AR in the Saha and Gangseo sea regions. The analysis indicated that ARs deployed in the Busan sea region maintained a good conservation ratio of more than 85%. The loss in AR volume (approximately 15%) was due to subsidence, breakaway, and incomplete deployment.

• Journal of Wetlands Research
• /
• v.17 no.1
• /
• pp.91-100
• /
• 2015

The aim of this study is to compare and analyze plant monitoring results of 2012 and 2013 for three artificial wetlands of Binae, Sedo, and Okpo areas in Namhan river, Keum river, and Nakdong river respectively. As the results, the Binae wetland in 2013 shows the same environment with 2012, that is, Willow and Common reed were dominant and the distributed plants in the wetland were Phragmites japonica Steud, Carex dispalta Boott, and Humulus japonicus in lowland of river side. Especially, Humulus japonicus was increased in summer season and Willow was increased in lowland of 초원. In Okpo wetland, Water chestnut was appeared in 2013 which it was not in 2012. Reed distribution was reduced in 2013 comparing to 2012 but Pussy willow distribution was not changed in 2013. In Sedo wetland, Secondary grassland was artificially planted in 2013 and so the colony was changed to Chinese Lespedeza and also most of plant colony was changed to Colt's-tail. Therefore, we can know that wetted transition rather than dried transition will be occurred in the Binae wetland. However, the Okpo wetland has monotonic change and so the transition will be proceeded with long time. The Sedo wetland showed wetted and dried transitions exist together. Therefore, the wetland will be changed to Reeds, Common reed, and Willow colonies.

• The Korean Journal of Ecology
• /
• v.17 no.3
• /
• pp.415-424
• /
• 1994

Community structure, size and age distribution of tree species of the subalpine Korean fir (Abies koreana) forest at the elevation of 1, 400~1, 700m were strudied in the north-west side of Chonwangbong Peak (1915m) in Mt. Chiri for the purpose of better understanding of the population dynamics of Abies koreana. Eight 20m X 20m permanent quadrats were established in 1991, and trees $\geq$ 2.5cm DBH for all species were marked with numbered aluminum tag, and saplings of Korean fir and Korean pine (Pinus koraiensis) were also tagged with aluminum sheet. These two conifer species comprised more than 60% of the total density and total basal area of the community. Quercus mongolica and Acer pseudo-sieboldianum were subdominats. Oridination study showed that cool temperate species such as Sorbus commixta, Betula costata, Acer tschonoskii and Acer ukurunduense occurred in close association with Abies koreana and Pinus koraiensis. Major tree species, especially Abies koreana and Pinus koraiensis, were well represented in samller size classes, indicating that they were regenerating well. Age distributuion of the Abies koreana and Pinus koraiensis showed that the former has longer physiological longevity than the latter, and that establishment were episodic and varied with sites, which implies the importance of the role of natural and artificial disturbances in this Abies kireana forest.

• The Korean Journal of Ecology
• /
• v.16 no.3
• /
• pp.305-316
• /
• 1993

Stands profiles, yearly changes in growth of annual rings, age and diameter structure, and spatial distribution pattern of individuals in the Pinus densiflora stands around the Yeocheon industrial complex were investigated. Growth of annual ring in Pinus densiflora, which survived when vegetation of this area was damaged by air pollutants, was suppressed for about 10 years since 1974 when factories in this area began to operate, but since then such suppressed growth tended to be recovered. It was supposed that the suppresed growth was originated from air pollution and that improvement of growth since the suppressed period was due to the release from competition with them by death of neighbouring trees and the resuction of the amount of air pollutants. Physiognomy of Pinus densiflora stands showed mosaic pattern composed of different patches. Spatial distribution pattern of individuals an stand profiles were similar to those of Pinus densiflora stands regenerated after natural and artificial disturbances. In an age distribution diagram, age of Pinus densiflora population ranged from 1 to 33 years, Among these individuals were recrited corresponded to the suppresed period of growth of annual ring in Pinus densiflora survived when the vegetation was damaged by air pollution. On the other hand, from the result of analysis of frequency distribution diagram of diameter, it was postulated that even if whis Pinus densiflora community can be maintained as it is for the time being, it might be changed to Quercus community with the lapse of time.

• Journal of Astronomy and Space Sciences
• /
• v.31 no.4
• /
• pp.285-294
• /
• 2014

The uneven mass distribution of the Moon highly perturbs the lunar spacecrafts. This uneven mass distribution leads to peculiar dynamical features of the lunar orbiters. The critical inclination is the value of inclination which keeps the deviation of the argument of pericentre from the initial values to be zero. Considerable investigations have been performed for critical inclination when the gravity field is assumed to be symmetric around the equator, namely for oblate gravity field to which Earth's satellites are most likely to be subjected. But in the case of a lunar orbiter, the gravity field of mass distribution is rather asymmetric, that is, sectorial, and tesseral, harmonic coefficients are big enough so they can't be neglected. In the present work, the effects of the first sectorial and tesseral harmonic coefficients in addition to the first zonal harmonic coefficients on the critical inclination of a lunar artificial satellite are investigated. The study is carried out using the Hamiltonian framework. The Hamiltonian of the problem is cconstructed and the short periodic terms are eliminated using Delaunay canonical variables. Considering the above perturbations, numerical simulations for a hypothetical lunar orbiter are presented. Finally, this study reveals that the critical inclination is quite different from the critical inclination of traditional sense and/or even has multiple solutions. Consequently, different families of critical inclination are obtained and analyzed.

• International Journal of Advanced Culture Technology
• /
• v.7 no.2
• /
• pp.209-217
• /
• 2019

Due to the development of artificial intelligence and image recognition technology that play important roles in the field of 4th industry, office automation systems and unmanned automation systems are rapidly spreading in human society. The proposed algorithm first finds the variances of the differences between the tile values constituting the learning characters and the experimental character and then recognizes the experimental character according to the distribution of the three learning characters with the smallest variances. In more detail, for 100 learning data characters and 10 experimental data characters, each character is defined as the number of black pixels belonging to 15 tile areas. For each character constituting the experimental data, the variance of the differences of the tile values of 100 learning data characters is obtained and then arranged in the ascending order. After that, three learning data characters with the minimum variance values are selected, and the final recognition result for the given experimental character is selected according to the distribution of these character types. Moreover, we compare the recognition result with the result made by a neural network of basic structure. It is confirmed that satisfactory recognition results are obtained through the processes that subdivide the learning characters and experiment characters into tile sizes and then select the recognition result using variances.

• KEPCO Journal on Electric Power and Energy
• /
• v.7 no.1
• /
• pp.115-123
• /
• 2021

Through machine learning-based load prediction, it is possible to prevent excessive power generation or unnecessary economic investment by estimating the appropriate amount of facility investment in consideration of the load that will increase in the future or providing basic data for policy establishment to distribute the maximum load. However, in order to secure the reliability of the developed load prediction model in the field, the performance comparison verification between the distribution line load prediction models must be preceded, but a comparative performance verification system between the distribution line load prediction models has not yet been established. As a result, it is not possible to accurately determine the performance excellence of the load prediction model because it is not possible to easily determine the likelihood between the load prediction models. In this paper, we developed a reliability verification system for load prediction models including a method of comparing and verifying the performance reliability between machine learning-based load prediction models that were not previously considered, verification process, and verification result visualization methods. Through the developed load prediction model reliability verification system, the objectivity of the load prediction model performance verification can be improved, and the field application utilization of an excellent load prediction model can be increased.