• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.5
• /
• pp.999-1008
• /
• 2023

Early prediction of chronic diseases such as diabetes is an important issue, and improving the accuracy of diabetes prediction is especially important. Various machine learning and deep learning-based methodologies are being introduced for diabetes prediction, but these technologies require large amounts of data for better performance than other methodologies, and the learning cost is high due to complex data models. In this study, we aim to verify the claim that DNN using the pima dataset and k-fold cross-validation reduces the efficiency of diabetes diagnosis models. Machine learning classification methods such as decision trees, SVM, random forests, logistic regression, KNN, and various ensemble techniques were used to determine which algorithm produces the best prediction results. After training and testing all classification models, the proposed system provided the best results on XGBoost classifier with ADASYN method, with accuracy of 81%, F1 coefficient of 0.81, and AUC of 0.84. Additionally, a domain adaptation method was implemented to demonstrate the versatility of the proposed system. An explainable AI approach using the LIME and SHAP frameworks was implemented to understand how the model predicts the final outcome.

• KIISE Transactions on Computing Practices
• /
• v.21 no.1
• /
• pp.46-51
• /
• 2015

The explosive growth of data and the rapidly changing technical social evolution new analysis paradigm for predicting and reacting the future the past and present ig data. Prescriptive analysis has a fundamental difference because can support specific behaviors and results according to user's goals with defin researchers establish judgments and activities achiev the goals. However research methods not widely implemented and even the terminology, Prescriptive analysis, is still unfamiliar. This paper thus propose an infrastructure in the prescriptive analysis field with key considerations for enhancing capability of researchers through a case study based on InSciTe Advisory developed with scientific big data. InSciTe Advisory system s developed in 2013, and offers a prescriptive analytics report which contains various As-Is analysis results and To-Be analysis results 5W1H methodology. InSciTe Advisory therefore shows possibility strategy aims to reach a target role model group. Through the availability and reliability of the measurement model the evaluation results obtained relative advantage of 118.8% compared to Elsevier SciVal.

• The Journal of the Korea Contents Association
• /
• v.15 no.12
• /
• pp.129-140
• /
• 2015

The objective of this study is mainly to examine factors affecting utilization of graphical emoticons in the environment of mobile messenger. For this purpose, this research identified several determinants including demographic variables that have influences on utilization of graphical emoticons by an overview of prior research on Big 5 model and Uses and Gratification (U & G). An online survey was employed to collect data, and hierarchical regression analysis was used for data analysis. The results showed that females and younger respondents have higher utilization of graphical emoticons than males and the older. The findings also showed that extraversion as personal traits has influences on the utilization. And they indicated that people increase their utilization of graphical emoticons when they want to communicate with others efficiently and succinctly, and to follow the trend. The practical and theoretical implications of the findings in this study are also discussed.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.31-34
• /
• 2024

태양광 에너지는 탄소 중립 이행을 위한 주요 방안으로 많은 주목을 받고 있다. 태양광 발전량은 여러 환경적 요인에 따라 크게 달라질 수 있으므로, 정확한 발전량 예측은 전력 네트워크의 안정성과 효율적인 에너지 관리에 근본적으로 중요하다. 대표적인 인공지능 기술인 신경망(Neural Network)은 불안정한 환경 변수와 복잡한 상호작용을 효과적으로 학습할 수 있어 태양광 발전량 예측에서 우수한 성능을 도출하였다. 하지만, 신경망은 모델의 구조나 초매개변수(Hyperparameter)를 최적화하는 것은 복잡하고 시간이 많이 드는 작업이므로, 에너지 분야에서 실제 산업 적용에 한계가 존재한다. 본 논문은 2단계 신경망 최적화를 통한 태양광 발전량 예측 기법을 제안한다. 먼저, 태양광 발전량 데이터 셋을 훈련 집합과 평가 집합으로 분할한다. 훈련 집합에서, 각기 다른 은닉층의 개수로 구성된 여러 신경망 모델을 구성하고, 모델별로 Optuna를 적용하여 최적의 초매개변숫값을 선정한다. 다음으로, 은닉층별 최적화된 신경망 모델을 이용해 훈련과 평가 집합에서는 각각 5겹 교차검증을 적용한 발전량 추정값과 예측값을 출력한다. 마지막으로, 스태킹 앙상블 방식을 채택해 기본 초매개변숫값으로 설정해도 우수한 성능을 도출하는 랜덤 포레스트를 이용하여 추정값을 학습하고, 평가 집합의 예측값을 입력으로 받아 최종 태양광 발전량을 예측한다. 인천 지역으로 실험한 결과, 제안한 방식은 모델링이 간편할 뿐만 아니라 여러 신경망 모델보다 우수한 예측 성능을 도출하였으며, 이를 바탕으로 국내 에너지 산업에 이바지할 수 있을 것으로 기대한다.

• 한국신재생에너지학회:학술대회논문집
• /
• 2011.11a
• /
• pp.33.2-33.2
• /
• 2011

The 'NREL Phase VI' model with a 10.06m diameter was tested in the NASA Ames tunnel to make a reference data of the computational models. The test was conducted at the one rotational speed, blade tip speed 38m/s and the Reynolds number of the sectional airfoils in that test was around 1E6. The 1/5 scale down model of the 'NREL Phase VI' model was used in this paper to study the power characteristics in low Reynolds number region, 0.1E6 ~ 0.4E6 which is achievable range for the conventional wind tunnel facilities. The torque generated by the blade was directly measured by using the torque sensor installed in the rotating axis for a given wind speed and rotational speed. The power characteristics below the stall condition, lambda > 4, was presented in this paper. The power coefficient is very low in the condition below the Re. 0.2E6 and rapidly increases as the Re. increases. And it still increases but the variation is not so big in the condition above the Re. 0.3E6. This results shows that to study the performance of the wind turbine blade by using the scaled down model, the Re. should be larger than the 0.3E6.

• Convergence Security Journal
• /
• v.23 no.1
• /
• pp.79-87
• /
• 2023

Precision Medicine, which utilizes personal health information, genetic information, clinical information, etc., is growing as the next-generation medical industry. In Korea, medical institutions and information communication companies have coll aborated to provide cloud-based Precision Medicine Hospital Information Systems (P-HIS) to about 90 primary medical ins titutions over the past five years, and plan to continue promoting and expanding it to primary and secondary medical insti tutions for the next four years. Precision medicine is directly related to human health and life, making information protecti on and healthcare information protection very important. Therefore, this paper analyzes the preliminary research on inform ation protection models that can be utilized in cloud-based Precision Medicine Hospital Information Systems and ultimately proposes research on ways to improve information protection in P-HIS.

• Journal of Digital Contents Society
• /
• v.19 no.4
• /
• pp.739-747
• /
• 2018

With a large feature space data, feature selection has become an extremely important procedure in the Data Mining process. But the traditional feature selection methods with single process may no longer fit for this procedure. In this paper, we proposed a hybrid efficient feature selection model for high dimensional data. We have applied our model on KNHNAES data set, the result shows that our model outperforms many existing methods in terms of accuracy over than at least 5%.

• Journal of Convergence for Information Technology
• /
• v.8 no.5
• /
• pp.107-113
• /
• 2018

With the rapid development of big data and Internet technologies among IT technologies, it is being changed into an environment where data stored in the cloud environment can be used wherever the Internet is connected, without storing important data in an external storage device such as USB. However, protection of users' privacy information is becoming increasingly important as the data being processed in the cloud environment is changed into an environment that can be easily handled. In this paper, we propose a user-reserving management model that can improve the user 's service quality without exposing the information used in the cloud environment to a third party. In the proposed model, user group is grouped into virtual environment so that third party can not handle user's privacy information among data processed in various cloud environments, and then identity property and access control policy are processed by block chain.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.42 no.2
• /
• pp.23-28
• /
• 2005

This study was to suggest the propagation path loss and predictive model of propagation path analysis in order to apply the frequency in the millimeter-wave band to the real time inter-vehicle communication system. This study was to suppose the case of inter-vehicle communication on the one-way two-lanes road in the big cites with a lot of traffic jams in order to analyze the effect by the reflected wave of multipath. As a simulation of suggested model, it found out that the propagation path by the reflected wave was about 0.1[m]$\sim$5.1[m] longer than the one by the direct wave during the transmission of 100[m] wave direct path. Also, as a result of comparing the propagation path loss, the loss would be about -0.8[dB]$\sim$-4.2[dB] larger in case of wall reflection and -0.8[dB]$\sim$-1[dB] vehicle reflection. From the result above, this researcher found out that the path loss of reflected wave produced by the walls was about -3.2[dB] larger than the path loss produced by the adjacent vehicles.

• The KIPS Transactions:PartB
• /
• v.12B no.5 s.101
• /
• pp.561-566
• /
• 2005

This paper presents a visual tracking algorithm for the golf swing motion analysis by using the information of the pixels of video frames and movement of the golf club to solve the problem fixed center point in model based tracking method. The model based tracking method use the polynomial function for trajectory displaying of upswing and downswing. Therefore it is under the hypothesis of the no movement of the center of gravity so this method is not for the amateurs. we proposed method using the information of pixel and movement, we first detected the motion by using the information of pixel in the frames in golf swing motion. Then we extracted the club head and hand by a properties of club shaft that consist of the parallel line and the moved location of club in up-swing and down-swing. In addition, we can extract the center point of user by tracking center point of the line between center of head and both foots. And we made an experiment with data that movement of center point is big. Finally, we can track the real trajectory of club head, hand and center point by using proposed tracking algorithm.