• Computers and Concrete
• /
• v.33 no.2
• /
• pp.137-145
• /
• 2024

The technique of experimentally determining concrete's compressive strength for a given mix design is time-consuming and difficult. The goal of the current work is to propose a best working predictive model based on different machine learning algorithms such as Gradient Boosting Machine (GBM), Stacked Ensemble (SE), Distributed Random Forest (DRF), Extremely Randomized Trees (XRT), Generalized Linear Model (GLM), and Deep Learning (DL) that can forecast the compressive strength of ternary geopolymer concrete mix without carrying out any experimental procedure. A geopolymer mix uses supplementary cementitious materials obtained as industrial by-products instead of cement. The input variables used for assessing the best machine learning algorithm not only include individual ingredient quantities, but molarity of the alkali activator and age of testing as well. Myriad statistical parameters used to measure the effectiveness of the models in forecasting the compressive strength of ternary geopolymer concrete mix, it has been found that GBM performs better than all other algorithms. A sensitivity analysis carried out towards the end of the study suggests that GBM model predicts results close to the experimental conditions with an accuracy between 95.6 % to 98.2 % for testing and training datasets.

• Journal of Korean Society of Forest Science
• /
• v.90 no.1
• /
• pp.133-138
• /
• 2001

Increased base cation loss and Al mobilization, a consequence of soil acid neutralization responses, are common in air polluted areas showing forest decline. The prediction models of acid neutralization responses were developed by using indicators of soil acidification level(pH, and base saturation) in order to assess the forest soil sensitivity to acidification. The soil acidification level was greatest in Namsan followed by Kanghwa, Ulsan, and Hongcheon, being contrary to regional total $ANC_H$ pattern through soil columns leached with additional acid ($16.7mmol_c\;H^+/kg$), Both base exchange and Al dissolution were main acid neutralization processes in all study regions. There were low base exchange and high Al dissolution in the regions of the low total $ANC_H$. The $ANC_M$ by sulfate adsorption was greatest in Hongcheon compared with other regions even though the AN rate was very low as 6.4%. Coefficients of adjusted determination of simple and multiple regression models between soil acidification level indicators and the acid neutralization responses were more than 0.52(p<0.04) and 0.89(p<0.01), respectively. The result suggests that soil pH and base saturation are available indicators for predicting the acid neutralization responses. These prediction models could be used as an useful method to measure forest soil sensitivity to acidification.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.5 no.4
• /
• pp.171-178
• /
• 1997

Vehicle interior noise has become increasingly important in this recent years. The noise of a vehicle is one of the important problems in a vehicle design. The interior noise is caused by various vibration sources of vehicle compartment. The booming noise of a vehicle can be significantly affected by vibrations transmitted from engine excitation forces to the vehicle body. Specially, we are interested in the state of transmission paths such as engine mounts to reduce noise in a vehicle compartment. In this paper, we have been calculated the contribution of each transmission path such as engine mounts to interior noise. To identify contribution of each input sources and transmission paths to output, the effectiveness of each input component to output is calculated. Sensitivity analysis is carried out for investigation of contribution to output due to input variations. With the simulation of magnitude and phase change of inputs using vector synthesis diagram, the trends of synthesized output vector are obtained. As a result, we suggested sensitivity analysis of vector synthesis as a technique of prediction and control for noise in a vehicle compartment.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.15 no.6
• /
• pp.2109-2124
• /
• 1991

A simulation program is developed to analyse the performance of an axial flow turbine stage based on the meanline prediction method. The gradient projection method is utilized to minimize the aerodynamic losses under the specified constraints on such as flow coefficient, total pressure ratio, stage power and blade loading coefficient. After obtaining the optimum point for minimizing the stage loss, a sensitivity analysis is carried out ground the optimum point to find the effects of the design variables and the design constraints on the stage performance. The result of the senitivity analysis under a constant blade loading coefficient shows that the total loss is more sensitive to the mean diameter, the absolute flow angle at nozzle outlet, the relative flow angle at rotor outlet and the axial mean velocity compared to the chords and the pitches. Moreover, the design constraints on the degree of reaction at root and the blade length-to-diameter ratio are found to be most influencial on the maximization of the overall aerodynamic efficiency.

• Atmosphere
• /
• v.17 no.4
• /
• pp.435-453
• /
• 2007

KEOP (Korea Enhanced Observing Period)-2004 intensive summer observation was carried out from 20 June to 5 July 2004 over the Southwestern part of the Korean peninsula. In this study, the effects of KEOP-2004 intensive observation data on the simulation of precipitation system are investigated using KLAPS (Korea Local Analysis and Prediction System) and PSU/NCAR MM5. Three precipitation cases during the intensive observation are selected for detailed analysis. In addition to the control experiments using the traditional data for its initial and boundary conditions, two sensitivity experiments using KEOP data with and without Jindo radar are performed. Although it is hard to find a clear and consistent improvement in the verification score (threat score), it is found that the KEOP data play a role in improving the position and intensity of the simulated precipitation system. The experiments started at 00 and 12 UTC show more positive effect than those of 06 and 18 UTC. The effect of Jindo radar is dependent on the case. It plays a significant role in the heavy rain cases related to a mesoscale low over Changma front and the landing of a Typhoon. KEOP data produce more strong difference in the 06/18 UTC experiments than in 00/12 UTC, but give more positive effects in 00/12 UTC experiments. One of the possible explanations for this is that : KEOP data could properly correct the atmosphere around them when there are certain amounts of data, while gives excessive effect to the atmospheric field when there are few data. CRA analysis supports this reasoning. According to the CRA (Contiguous Rain Area) analysis, KEOP data in 00/12 UTC experiments improve only the surrounding area, resulting in essentially same precipitation system so the effects remain only in each convective cell rather than the system itself. On the other hand, KEOP data modify the precipitation system itself in 06/18 UTC experiments. Therefore the effects become amplified with time integration.

• International Journal of Aeronautical and Space Sciences
• /
• v.8 no.1
• /
• pp.95-104
• /
• 2007

A new design approach of complex geometries such as wing/body configuration is arranged by using overset mesh techniques under large scale computing environment. For an in-depth study of the flow physics and highly accurate design, several special overlapped structured blocks such as collar grid, tip-cap grid, and etc. which are commonly used in refined drag prediction are adopted to consider the applicability of the present design tools to practical problems. Various pre- and post-processing techniques for overset flow analysis and sensitivity analysis are devised or implemented to resolve overset mesh techniques into the design optimization problem based on Gradient Based Optimization Method (GBOM). In the pre-processing, the convergence characteristics of the flow solver and sensitivity analysis are improved by overlap optimization method. Moreover, a new post-processing method, Spline-Boundary Intersecting Grid (S-BIG) scheme, is proposed by considering the ratio of cell area for more refined prediction of aerodynamic coefficients and efficient evaluation of their sensitivities under parallel computing environment. With respect to the sensitivity analysis, discrete adjoint formulations for overset boundary conditions are derived by a full hand-differentiation. A smooth geometric modification on the overlapped surface boundaries and evaluation of grid sensitivities can be performed by mapping from planform coordinate to the surface meshes with Hicks-Henne function. Careful design works for the drag minimization problems of a transonic wing and a wing/body configuration are performed by using the newly-developed and -applied overset mesh techniques. The results from design applications demonstrate the capability of the present design approach successfully.

• Geomechanics and Engineering
• /
• v.12 no.3
• /
• pp.441-464
• /
• 2017

The determination of the mixture parameters of stabilization has become a great concern in geotechnical applications. This paper presents an effort about the application of artificial intelligence (AI) techniques including radial basis neural network (RBNN), multi-layer perceptrons (MLP), generalized regression neural network (GRNN) and adaptive neuro-fuzzy inference system (ANFIS) in order to predict the unconfined compressive strength (UCS) of silty soil stabilized with bottom ash (BA), jute fiber (JF) and steel fiber (SF) under different freeze-thaw cycles (FTC). The dosages of the stabilizers and number of freeze-thaw cycles were employed as input (predictor) variables and the UCS values as output variable. For understanding the dominant parameter of the predictor variables on the UCS of stabilized soil, a sensitivity analysis has also been performed. The performance measures of root mean square error (RMSE), mean absolute error (MAE) and determination coefficient ($R^2$) were used for the evaluations of the prediction accuracy and applicability of the employed models. The results indicate that the predictions due to all AI techniques employed are significantly correlated with the measured UCS ($p{\leq}0.05$). They also perform better predictions than nonlinear regression (NLR) in terms of the performance measures. It is found from the model performances that RBNN approach within AI techniques yields the highest satisfactory results (RMSE = 55.4 kPa, MAE = 45.1 kPa, and $R^2=0.988$). The sensitivity analysis demonstrates that the JF inclusion within the input predictors is the most effective parameter on the UCS responses, followed by FTC.

• Journal of Korean Academy of Nursing
• /
• v.50 no.4
• /
• pp.621-630
• /
• 2020

Purpose: This study was aimed to evaluate the external validity of a carbapenem-resistant Enterobacteriaceae (CRE) acquisition risk prediction model (the CREP-model) in a medium-sized hospital. Methods: This retrospective cohort study included 613 patients (CRE group: 69, no-CRE group: 544) admitted to the intensive care units of a 453-beds secondary referral general hospital from March 1, 2017 to September 30, 2019 in South Korea. The performance of the CREP-model was analyzed with calibration, discrimination, and clinical usefulness. Results: The results showed that those higher in age had lower presence of multidrug resistant organisms (MDROs), cephalosporin use ≥ 15 days, Acute Physiology and Chronic Health Evaluation II (APACHE II) score ≥ 21 points, and lower CRE acquisition rates than those of CREP-model development subjects. The calibration-in-the-large was 0.12 (95% CI: - 0.16~0.39), while the calibration slope was 0.87 (95% CI: 0.63~1.12), and the concordance statistic was .71 (95% CI: .63~.78). At the predicted risk of .10, the sensitivity, specificity, and correct classification rates were 43.5%, 84.2%, and 79.6%, respectively. The net true positive according to the CREP-model were 3 per 100 subjects. After adjusting the predictors' cutting points, the concordance statistic increased to .84 (95% CI: .79~.89), and the sensitivity and net true positive was improved to 75.4%. and 6 per 100 subjects, respectively. Conclusion: The CREP-model's discrimination and clinical usefulness are low in a medium sized general hospital but are improved after adjusting for the predictors. Therefore, we suggest that institutions should only use the CREP-model after assessing the distribution of the predictors and adjusting their cutting points.

• Computers and Concrete
• /
• v.31 no.3
• /
• pp.253-265
• /
• 2023

In this research, the gene expression programming (GEP) technique was employed to provide a new model for predicting the maximum loading capacity of concrete-encased steel (CES) columns. This model was developed based on 96 CES column specimens available in the literature. The six main parameters used in the model were the compressive strength of concrete (f_c), yield stress of structural steel (f_ys), yield stress of steel rebar (f_yr), and cross-sectional areas of concrete, structural steel, and steel rebar (A_c, A_s and A_r respectively). The performance of the prediction model for the ultimate load-carrying capacity was investigated using different statistical indicators such as root mean square error (RMSE), correlation coefficient (R), mean absolute error (MAE), and relative square error (RSE), the corresponding values of which for the proposed model were 620.28, 0.99, 411.8, and 0.01, respectively. Here, the predictions of the model and those of available codes including ACI ITG, AS 3600, CSA-A23, EN 1994, JGJ 138, and NZS 3101 were compared for further model assessment. The obtained results showed that the proposed model had the highest correlation with the experimental data and the lowest error. In addition, to see if the developed model matched engineering realities and corresponded to the previously developed models, a parametric study and sensitivity analysis were carried out. The sensitivity analysis results indicated that the concrete cross-sectional area (A_c) has the greatest effect on the model, while parameter (f_yr) has a negligible effect.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.125-140
• /
• 2013

We call a data set in which the number of records belonging to a certain class far outnumbers the number of records belonging to the other class, 'imbalanced data set'. Most of the classification techniques perform poorly on imbalanced data sets. When we evaluate the performance of a certain classification technique, we need to measure not only 'accuracy' but also 'sensitivity' and 'specificity'. In a customer churn prediction problem, 'retention' records account for the majority class, and 'churn' records account for the minority class. Sensitivity measures the proportion of actual retentions which are correctly identified as such. Specificity measures the proportion of churns which are correctly identified as such. The poor performance of the classification techniques on imbalanced data sets is due to the low value of specificity. Many previous researches on imbalanced data sets employed 'oversampling' technique where members of the minority class are sampled more than those of the majority class in order to make a relatively balanced data set. When a classification model is constructed using this oversampled balanced data set, specificity can be improved but sensitivity will be decreased. In this research, we developed a hybrid model of support vector machine (SVM), artificial neural network (ANN) and decision tree, that improves specificity while maintaining sensitivity. We named this hybrid model 'hybrid SVM model.' The process of construction and prediction of our hybrid SVM model is as follows. By oversampling from the original imbalanced data set, a balanced data set is prepared. SVM_I model and ANN_I model are constructed using the imbalanced data set, and SVM_B model is constructed using the balanced data set. SVM_I model is superior in sensitivity and SVM_B model is superior in specificity. For a record on which both SVM_I model and SVM_B model make the same prediction, that prediction becomes the final solution. If they make different prediction, the final solution is determined by the discrimination rules obtained by ANN and decision tree. For a record on which SVM_I model and SVM_B model make different predictions, a decision tree model is constructed using ANN_I output value as input and actual retention or churn as target. We obtained the following two discrimination rules: 'IF ANN_I output value <0.285, THEN Final Solution = Retention' and 'IF ANN_I output value ${\geq}0.285$, THEN Final Solution = Churn.' The threshold 0.285 is the value optimized for the data used in this research. The result we present in this research is the structure or framework of our hybrid SVM model, not a specific threshold value such as 0.285. Therefore, the threshold value in the above discrimination rules can be changed to any value depending on the data. In order to evaluate the performance of our hybrid SVM model, we used the 'churn data set' in UCI Machine Learning Repository, that consists of 85% retention customers and 15% churn customers. Accuracy of the hybrid SVM model is 91.08% that is better than that of SVM_I model or SVM_B model. The points worth noticing here are its sensitivity, 95.02%, and specificity, 69.24%. The sensitivity of SVM_I model is 94.65%, and the specificity of SVM_B model is 67.00%. Therefore the hybrid SVM model developed in this research improves the specificity of SVM_B model while maintaining the sensitivity of SVM_I model.