• 환경영향평가
• /
• 제29권4호
• /
• pp.272-285
• /
• 2020

본 연구에서는 3차원 기상 및 대기질 모델의 입출력 자료를 평가하는 데 필요한 통계 검증지표를 선별하고, 선정된 검증지표의 기준치를 조사하여 그 결과를 요약하였다. 여러 국내외 문헌과 최근 논문 검토를 통해 최종 선정된 통계 검증지표는 MB (Mean Bias), ME (Mean Error), MNB (Mean Normalized Bias Error), MNE (Mean Absolute Gross Error), RMSE (Root Mean Square Error), IOA (Index of Agreement), R (Correlation Coefficient), FE (Fractional Error), FB (Fractional Bias)로 총 9가지이며, 국내외 문헌을 통해 그 기준치를 확인하였다. 그 결과, 기상모델의 경우 대부분 MB와 ME가 주요 지표로 사용되어 왔고, 대기질 모델 결과는 NMB와 NME 지표가 주로 사용되었으며, 그 기준치의 차이를 분석하였다. 아울러 이들 통계 검증지표값을 이용하여 모델 예측 결과를 효과적으로 비교하기 위한 표출 도식으로 축구 도식, 테일러 도식, Q-Q (Quantile-Quantile) 도식의 장단점을 분석하였다. 나아가 본 연구 결과를 기반으로 우리나라의 산악지역의 특수성 등이 잘 고려된 통계 검증지표의 기준치 설정 등의 추가연구가 효과적으로 진행될 수 있기를 기대한다.

• 한국통신학회논문지
• /
• 제22권10호
• /
• pp.2303-2309
• /
• 1997

본 논문에서는 고차통계에 의한 적응알고리즘 가운데 오차의 평균절대값삼승(LMAT)을 최소화하는 알고리즘의 수렴특성에 대하여 분석하였다. 사용된 입력신호가 Gaussian 분포를 갖는다는 가정하에 알고리즘의 평균자승 추정오차와 필터계수의 평균 및 평균자승 특성에 대해 정량적인 분석을 수행하였으며, 이에 대한 관계식을 각각 유도하였다. 이론적으로 분석된 결과는 컴퓨터 모의실험에 의하여 그 타당성을 검증하였고, 이론치와 실험치가 거의 일치함을 확인하였다.

• 한국전자통신학회논문지
• /
• 제16권4호
• /
• pp.625-632
• /
• 2021

본 논문은 잡음감쇠기에서 CNN(Convolutional Neural Network) 계층의 필터 수가 성능에 미치는 영향을 연구하였다 이 시스템은 적응필터 대신 신경망 예측필터를 이용하며 심층학습방법으로 잡음을 감쇠한다. 64-뉴런, 16-커널 CNN 필터와 오차 역전파 알고리즘을 이용하여 잡음이 포함된 음성신호로부터 음성을 추정한다. 본 연구에서 필터 수에 대한 잡음감쇠기의 성능을 검증하기 위하여 Keras 라이브러리를 사용한 프로그램을 작성하고 시뮬레이션을 실시하였다. 시뮬레이션 결과, 본 시스템은 필터 수가 16일 때 MSE(Mean Squared Error) 및 MAE(Mean Absolute Error) 값이 가장 작은 것으로 나타났으며 필터가 4개 일 때 성능이 가장 낮은 것을 볼 수 있다. 그리고 필터가 8개 이상이 되면 필터 수에 따라 MSE 및 MAE 값이 크게 차이나지 않는 것을 보여주었다. 이러한 결과로부터 음성신호의 주요 특징을 표현하기 위해서는 약 8개 이상의 필터를 사용해야 한다는 것을 알 수 있다.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2019년도 추계학술발표대회
• /
• pp.585-588
• /
• 2019

Predictions of Energy Consumption for Industries gain an important place in energy management and control system, as there are dynamic and seasonal changes in the demand and supply of energy. This paper presents and discusses the predictive models for energy consumption of the steel industry. Data used includes lagging and leading current reactive power, lagging and leading current power factor, carbon dioxide (tCO2) emission and load type. In the test set, four statistical models are trained and evaluated: (a) Linear regression (LR), (b) Support Vector Machine with radial kernel (SVM RBF), (c) Gradient Boosting Machine (GBM), (d) random forest (RF). Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) are used to measure the prediction efficiency of regression designs. When using all the predictors, the best model RF can provide RMSE value 7.33 in the test set.

• 한국멀티미디어학회논문지
• /
• 제25권2호
• /
• pp.440-449
• /
• 2022

In recent years, the air pollution and Air Quality Index (AQI) has been a pivotal point for researchers due to its effect on human health. Various research has been done in predicting the AQI but most of these studies, either lack dense temporal data or cover one or two air pollutant elements. In this paper, a hybrid Convolutional Neural approach integrated with recurrent neural network architecture (CNN-LSTM), is presented to find air pollution inference using a multivariate air pollutant elements dataset. The aim of this research is to design a robust and real-time air pollutant forecasting system by exploiting a neural network. The proposed approach is implemented on a 24-month dataset from Seoul, Republic of Korea. The predicted results are cross-validated with the real dataset and compared with the state-of-the-art techniques to evaluate its robustness and performance. The proposed model outperforms SVM, SVM-Polynomial, ANN, and RF models with 60.17%, 68.99%, 14.6%, and 6.29%, respectively. The model performs SVM and SVM-Polynomial in predicting O3 by 78.04% and 83.79%, respectively. Overall performance of the model is measured in terms of Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and the Root Mean Square Error (RMSE).

• Nuclear Engineering and Technology
• /
• 제55권4호
• /
• pp.1342-1349
• /
• 2023

The planning accuracy of charged particle therapy (CPT) is subject to the accuracy of stopping power (SP) estimation. In this study, we propose a method of deriving a pseudo-triple-energy CT (pTECT) that can be achievable in the existing dual-energy CT (DECT) systems for better SP estimation. In order to remove the direct effect of errors in CT values, relative CT values according to three scanning voltage settings were used. CT values of each tissue substitute phantom were measured to show the non-linearity of the values thereby suggesting the absolute difference and ratio of CT values as parameters for SP estimation. Electron density, effective atomic number (EAN), mean excitation energy and SP were calculated based on these parameters. Two of conventional methods were implemented and compared to the proposed pTECT method in terms of residuals, absolute error and root-mean-square-error (RMSE). The proposed method outperformed the comparison methods in every evaluation metrics. Especially, the estimation error for EAN and mean excitation using pTECT were converging to zero. In this proof-of-concept study, we showed the feasibility of using three CT values for accurate SP estimation. Our suggested pTECT method indicates potential clinical utility of spectral CT imaging for CPT planning.

• Structural Engineering and Mechanics
• /
• 제86권5호
• /
• pp.621-633
• /
• 2023

Stability analysis and support system estimation of the Beheshtabad water transmission tunnel is investigated in this research. A combination approach based on the rock mass rating (RMR) and rock mass quality index (Q) is used for this purpose. In the first step, 40 datasets related to the petrological, structural, hydrological, physical, and mechanical properties of tunnel host rocks are measured in the field and laboratory. Then, RMR, Q, and height of influenced zone above the tunnel roof are computed and sorted into five general groups to analyze the tunnel stability and determine its support system. Accordingly, tunnel stand-up time, rock load, and required support system are estimated for five sorted rock groups. In addition, various empirical relations between RMR and Q i.e., linear, exponential, logarithmic, and power functions are developed using the analysis of variance (ANOVA). Based on the significance level (sig.), determination coefficient (R²) and Fisher-test (F) indices, power and logarithmic equations are proposed as the optimum relations between RMR and Q. To validate the proposed relations, their results are compared with the results of previous similar equations by using the variance account for (VAF), root mean square error (RMSE), mean absolute percentage error (MAPE) and mean absolute error (MAE) indices. Comparison results showed that the accuracy of proposed RMR-Q relations is better than the previous similar relations and their outputs are more consistent with actual data. Therefore, they can be practically utilized in designing the tunneling projects with an acceptable level of accuracy and reliability.

• Advances in Energy Research
• /
• 제8권2호
• /
• pp.111-123
• /
• 2022

Solar Energy is the energy of solar radiation carried by them in the form of heat and light. It can be converted into electricity. Solar potential depends on the site's atmosphere; the solar energy distribution depends on many factors, e.g., turbidity, cloud types, pollution levels, solar altitude, etc. We estimated solar radiation with the help of the Ashrae clear-sky model for three locations in Pakistan, namely Pasni, Gwadar, and Jiwani. As these locations are close to each other as compared to the distance between the sun and earth, therefore a slight change of latitude and longitude does not make any difference in the calculation of direct beam solar radiation (BSR), diffuse solar radiation (DSR), and global solar radiation (GSR). A modified formula for declination angle is also developed and presented. We also created two different models for Ashrae constants. The values of these constants are compared with the standard Ashrae Model. A good agreement is observed when we used these constants to calculate BSR, DSR, GSR, the Root mean square error (RMSE), Mean Absolute error (MABE), Mean Absolute percent error (MAPE), and chisquare (χ²) values are in acceptance range, indicating the validity of the models.

• 한국정보통신학회논문지
• /
• 제25권7호
• /
• pp.884-889
• /
• 2021

국내 코로나19의 감염자 수가 백신과 사회적 거리 두기, 백신 등 여러 가지 노력 덕분에 차츰 줄어드는 듯 보였으나 2020년 2월 20일 특정한 사건 이후 감염자 수가 증가한 것처럼, 2020년 12월부터 또다시 급격히 감염자 수가 증가하는 추세이며 꾸준히 일일 500명가량의 감염자 수가 이어지고 있다. 따라서 Kaggle의 데이터셋을 이용해서 Prophet 알고리즘을 통해 미래 코로나19를 예측하고 사이킷런을 통해 결정계수, 평균 절대 오차, 평균 백분율 오차, 평균 제곱 차, 평균 제곱근 편차를 통해 이 예측에 대한 설명력을 더한다. 또한 코로나19가 급격히 특정한 사건이 없었을 경우 국내 감염자 수를 예측해 앞으로 우리가 미래의 질병에 대해서 방역과 방역 수칙 실천의 중요함을 강조한다.

• Communications for Statistical Applications and Methods
• /
• 제22권6호
• /
• pp.675-683
• /
• 2015

In the 21st century, we now face the serious problems of the enormous consumption of the energy resources. Depending on the power consumption increases, both China and South Korea face a reduction in available resources. This paper considers the regression models and time-series models to compare the performance of the forecasting accuracy based on Mean Absolute Percentage Error (MAPE) in order to forecast the electricity demand accurately on the short-term period (68 months) data in Northeast China and find the relationship with Korea. Among the models the support vector regression (SVR) model shows superior performance than time-series models for the short-term period data and the time-series models show similar results with the SVR model when we use long-term period data.