• 대한의용생체공학회:의공학회지
• /
• 제40권2호
• /
• pp.62-67
• /
• 2019

The purpose of this study was to evaluate the performance of deep neural network model in order to determine whether there is a risk factor for coronary artery disease based on the cardiac variation parameter. The study used unidentifiable 297 data to evaluate the performance of the model. Input data consists of heart rate parameters, which are SDNN (standard deviation of the N-N intervals), PSI (physical stress index), TP (total power), VLF (very low frequency), LF (low frequency), HF (high frequency), RMSSD (root mean square of successive difference) APEN (approximate entropy) and SRD (successive R-R interval difference), the age group and sex. Output data are divided into normal and patient groups, and the patient group consists of those diagnosed with diabetes, high blood pressure, and hyperlipidemia among the various risk factors that can cause coronary artery disease. Based on this, a binary classification model was applied using Deep Neural Network of deep learning techniques to classify normal and patient groups efficiently. To evaluate the effectiveness of the model used in this study, Kernel SVM (support vector machine), one of the classification models in machine learning, was compared and evaluated using same data. The results showed that the accuracy of the proposed deep neural network was train set 91.79% and test set 85.56% and the specificity was 87.04% and the sensitivity was 83.33% from the point of diagnosis. These results suggest that deep learning is more efficient when classifying these medical data because the train set accuracy in the deep neural network was 7.73% higher than the comparative model Kernel SVM.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제8권1호
• /
• pp.19-26
• /
• 2019

본 논문에서는 기타 타브 악보에서 추출한 프렛 번호를 대상으로 학습 알고리즘의 분류 성능을 비교한다. 타브 악보로부터 세그먼트를 통해 추출된 타브 숫자 데이터는 타브 선과 악보 기호가 포함하기 때문에 레이블링 기법과 비선형 필터를 이용하여 프렛 숫자를 추출한다. 추가적인 데이터 확보를 위해 전처리가 수행된 데이터에 대해 4 방향으로 이동 연산을 수행한다. 선택된 학습 모델은 베이지안 분류기, 지지벡터기기, 프로토타입 기반 학습, 다층 신경망 그리고 합성곱 신경망 모델 등이다. 실험 결과 베이지안 분류기는 85.0% 평균 정확도를 보였고 나머지 분류기는 99.0% 이상의 평균 정확도를 보였다. 일반화 성능과 전처리 단계를 고려 시 합성곱 신경망이 다른 학습 모델들보다 우수하다.

• 한국정보통신학회논문지
• /
• 제23권10호
• /
• pp.1290-1297
• /
• 2019

본 논문에서는 ultra-wideband(UWB) 시스템에서 합성곱 신경망(CNN)을 이용한 거리 추정 기법을 제안한다. 제안하는 기법은 UWB 신호를 이용하여 송신기와 수신기 사이의 거리를 추정하기 위하여 수신신호의 크기 샘플로 이루어진 1차원 벡터를 2차원 행렬로 재구성하며, 이 2차원 행렬로부터 합성곱 신경망 회귀를 이용하여 거리를 추정한다. IEEE 802.15.4a 표준의 UWB 실내 가시선 채널모델을 이용하여 수신신호를 생성하여 학습데이터를 만들며 합성곱 신경망 모델을 학습시킨다. 또한 실제 필드 시험을 통해 실내환경에서의 실험 데이터를 이용하여 거리추정 성능을 확인한다. 제안하는 기법은 기존의 문턱값 기반의 거리 추정 기법과의 성능비교도 수행하는데, 결과에 따르면 10m 거리에서 제안기법은 0.6m의 제곱근 평균 자승 에러를 보이는데 기존기법은 1.6m로 훨씬 큰 에러를 보인다.

• 한국지구과학회지
• /
• 제42권6호
• /
• pp.605-622
• /
• 2021

Seasonal forecasting has numerous socioeconomic benefits because it can be used for disaster mitigation. Therefore, it is necessary to diagnose and improve the seasonal forecast model. Moreover, the model performance is partly related to the ocean model. This study evaluated the hindcast performance in the upper ocean of the Global Seasonal Forecasting System version 5-Global Couple Configuration 2 (GloSea5-GC2) using a multivariable integrated evaluation method. The normalized potential temperature, salinity, zonal and meridional currents, and sea surface height anomalies were evaluated. Model performance was affected by the target month and was found to be better in the Pacific than in the Atlantic. An increase in lead time led to a decrease in overall model performance, along with decreases in interannual variability, pattern similarity, and root mean square vector deviation. Improving the performance for ocean currents is a more critical than enhancing the performance for other evaluated variables. The tropical Pacific showed the best accuracy in the surface layer, but a spring predictability barrier was present. At the depth of 301 m, the north Pacific and tropical Atlantic exhibited the best and worst accuracies, respectively. These findings provide fundamental evidence for the ocean forecasting performance of GloSea5.

• 한국추진공학회:학술대회논문집
• /
• 한국추진공학회 2008년 영문 학술대회
• /
• pp.790-798
• /
• 2008

Counter-rotating axial flow fan(CRF) consists of two counter-rotating rotors without stator blades. CRF shows the complex flow characteristics of the three-dimensional, viscous, and unsteady flow fields. For the understanding of the entire core flow in CRF, it is necessary to investigate the three-dimensional unsteady flow field between the rotors. This information is also essential to improve the aerodynamic characteristics and to reduce the aerodynamic noise level and vibration characteristics of the CRF. In this paper, experimental study on the three-dimensional unsteady flow of the CRF is performed at the design point(operating point). Flow fields in the CRF are measured at the cross-sectional planes of the upstream and downstream of each rotor using the $45^{\circ}$ inclined hot-wire. The phase-locked averaged hot-wire technique utilizes the inclined hot-wire, which rotates successively with 120 degree increments about its own axis. Three-dimensional unsteady flow characteristics such as tip vortex, secondary flow and tip leakage flow in the CRF are shown in the form of the axial, radial and tangential velocity vector plot and velocity contour. The phase-locked averaged velocity profiles of the CRF are analyzed by means of the stationary unsteady measurement technique. At the mean radius of the front rotor inlet and the outlet, the phase-locked averaged velocity profiles show more the periodical flow characteristics than those of the hub region. At the tip region of the CRF, the axial velocity is decreased due to the boundary layer effect of the fan casing and the tip vortex flow. The radial and the tangential velocity profiles show the most unstable and unsteady flow characteristics compared with other position of rotors. But, the phase-locked averaged velocity profiles of the downstream of the rear rotor show the aperiodic flow pattern due to the mixture of the front rotor wake period and the rear rotor rotational period.

• Asian Journal of Business Environment
• /
• 제13권2호
• /
• pp.9-20
• /
• 2023

Purpose: This study aims to map the spatial distribution of poverty using nighttime light satellite images as a proxy indicator of economic activities and infrastructure distribution in D.I Yogyakarta, Indonesia. Research design, data, and methodology: This study uses official poverty statistics (National Socio-economic Survey (SUSENAS) and Poverty Database 2015) to compare satellite imagery's ability to identify poor urban areas in D.I Yogyakarta. National Socioeconomic Survey (SUSENAS), as poverty statistics at the macro level, uses expenditure to determine the poor in a region. Poverty Database 2015 (BDT 2015), as poverty statistics at the micro-level, uses asset ownership to determine the poor population in an area. Pearson correlation is used to identify the correlation among variables and construct a Support Vector Regression (SVR) model to estimate the poverty level at a granular level of 1 km x 1 km. Results: It is found that macro poverty level and moderate annual nighttime light intensity have a Pearson correlation of 74 percent. It is more significant than micro poverty, with the Pearson correlation being 49 percent in 2015. The SVR prediction model can achieve the root mean squared error (RMSE) of up to 8.48 percent on SUSENAS 2020 poverty data.Conclusion: Nighttime light satellite imagery data has potential benefits as alternative data to support regional poverty mapping, especially in urban areas. Using satellite imagery data is better at predicting regional poverty based on expenditure than asset ownership at the micro-level. Light intensity at night can better describe the use of electricity consumption for economic activities at night, which is captured in spending on electricity financing compared to asset ownership.

• 예술인문사회 융합 멀티미디어 논문지
• /
• 제8권3호
• /
• pp.57-67
• /
• 2018

앙상블 기법은 기계학습에서 다수의 알고리즘을 사용하여 더 좋은 성능을 내기 위해 사용하는 방법이다. 본 논문에서는 앙상블 기법에서 많이 사용되는 부스팅과 배깅에 대해 소개를 하고, 서포트벡터 회귀, 방사기저함수 네트워크, 가우시안 프로세스, 다층 퍼셉트론을 이용하여 설계한다. 추가적으로 순환신경망과 MOHID 수치모델을 추가하여 실험을 진행한다. 실험적 검증를 위해 사용하는 뜰개 데이터는 7 개의 지역에서 관측된 683 개의 관측 자료다. 뜰개 관측 자료를 이용하여 6 개의 알고리즘과의 비교를 통해 앙상블 기법의 성능을 검증한다. 검증 방법으로는 평균절대오차를 사용한다. 실험 방법은 배깅, 부스팅, 기계학습을 이용한 앙상블 모델을 이용하여 진행한다. 각 앙상블 모델마다 동일한 가중치를 부여한 방법, 차등한 가중치를 부여한 방법을 이용하여 오류율을 계산한다. 가장 좋은 오류율을 나타낸 방법은 기계학습을 이용한 앙상블 모델로서 6 개의 기계학습의 평균에 비해 61.7%가 개선된 결과를 보였다.

• 국제학술발표논문집
• /
• The 8th International Conference on Construction Engineering and Project Management
• /
• pp.463-481
• /
• 2020

Cutter suction dredgers (CSDs) are widely used in various dredging constructions such as channel excavation, wharf construction, and reef construction. During a CSD construction, the main operation is to control the swing speed of cutter to keep the slurry concentration in a proper range. However, the slurry concentration cannot be monitored in real-time, i.e., there is a "time-lag effect" in the log of slurry concentration, making it difficult for operators to make the optimal decision on controlling. Concerning this issue, a solution scheme that using real-time monitored indicators to predict current slurry concentration is proposed in this research. The characteristics of the CSD monitoring data are first studied, and a set of preprocessing methods are presented. Then we put forward the concept of "index class" to select the important indices. Finally, an ensemble learning algorithm is set up to fit the relationship between the slurry concentration and the indices of the index classes. In the experiment, log data over seven days of a practical dredging construction is collected. For comparison, the Deep Neural Network (DNN), Long Short Time Memory (LSTM), Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), and the Bayesian Ridge algorithm are tried. The results show that our method has the best performance with an R² of 0.886 and a mean square error (MSE) of 5.538. This research provides an effective way for real-time predicting the slurry concentration of CSDs and can help to improve the stationarity and production efficiency of dredging construction.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권10호
• /
• pp.2643-2657
• /
• 2023

Scene graphs are structured representations that can clearly convey objects and the relationships between them, but are often heavily biased due to the highly skewed, long-tailed relational labeling in the dataset. Indeed, the visual world itself and its descriptions are biased. Therefore, Unbiased Scene Graph Generation (USGG) prefers to train models to eliminate long-tail effects as much as possible, rather than altering the dataset directly. To this end, we propose Geometric and Semantic Improvement (GSI) for USGG to mitigate this issue. First, to fully exploit the feature information in the images, geometric dimension and semantic dimension enhancement modules are designed. The geometric module is designed from the perspective that the position information between neighboring object pairs will affect each other, which can improve the recall rate of the overall relationship in the dataset. The semantic module further processes the embedded word vector, which can enhance the acquisition of semantic information. Then, to improve the recall rate of the tail data, the Class Balanced Seesaw Loss (CBSLoss) is designed for the tail data. The recall rate of the prediction is improved by penalizing the body or tail relations that are judged incorrectly in the dataset. The experimental findings demonstrate that the GSI method performs better than mainstream models in terms of the mean Recall@K (mR@K) metric in three tasks. The long-tailed imbalance in the Visual Genome 150 (VG150) dataset is addressed better using the GSI method than by most of the existing methods.

• International Journal of Computer Science & Network Security
• /
• 제24권2호
• /
• pp.101-112
• /
• 2024

Telugu language is considered as fourth most used language in India especially in the regions of Andhra Pradesh, Telangana, Karnataka etc. In international recognized countries also, Telugu is widely growing spoken language. This language comprises of different dependent and independent vowels, consonants and digits. In this aspect, the enhancement of Telugu Handwritten Character Recognition (HCR) has not been propagated. HCR is a neural network technique of converting a documented image to edited text one which can be used for many other applications. This reduces time and effort without starting over from the beginning every time. In this work, a Unicode based Handwritten Character Recognition(U-HCR) is developed for translating the handwritten Telugu characters into English language. With the use of Centre of Gravity (CG) in our model we can easily divide a compound character into individual character with the help of Unicode values. For training this model, we have used both online and offline Telugu character datasets. To extract the features in the scanned image we used convolutional neural network along with Machine Learning classifiers like Random Forest and Support Vector Machine. Stochastic Gradient Descent (SGD), Root Mean Square Propagation (RMS-P) and Adaptative Moment Estimation (ADAM)optimizers are used in this work to enhance the performance of U-HCR and to reduce the loss function value. This loss value reduction can be possible with optimizers by using CNN. In both online and offline datasets, proposed model showed promising results by maintaining the accuracies with 90.28% for SGD, 96.97% for RMS-P and 93.57% for ADAM respectively.