• Geophysics and Geophysical Exploration
• /
• v.25 no.4
• /
• pp.242-251
• /
• 2022

Automatic differentiation automatically calculates the derivatives of a function using the chain rule once the forward operation of a function is defined. Given the recent development of computing libraries that support automatic differentiation, many researchers have adopted automatic differentiation techniques to solve geophysical inverse problems. We analyzed the advantages, disadvantages, and performances of automatic differentiation techniques using the gradient calculations of seismic full waveform inversion objective functions. The gradients of objective functions can be expressed as multiplications of the derivatives of the model parameters, wavefields, and objective functions using the chain rule. Using numerical examples, we demonstrated the speed of analytic differentiation and the convenience of complex gradient calculations for automatic differentiation. We calculated derivatives of model parameters and objective functions using automatic differentiation and derivatives of wavefields using analytic differentiation.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.150-150
• /
• 2022

This study investigates the roles of decomposition methods on high accuracy in daily rainfall prediction from light gradient boosting machine (LightGBM) algorithm. Here, empirical mode decomposition (EMD) and singular spectrum analysis (SSA) methods were considered to decompose and reconstruct input time series into trend terms, fluctuating terms, and noise components. The decomposed time series from EMD and SSA methods were used as input data for LightGBM algorithm in two hybrid models, including empirical mode-based light gradient boosting machine (EMDGBM) and singular spectrum analysis-based light gradient boosting machine (SSAGBM), respectively. A total of four parameters (i.e., temperature, humidity, wind speed, and rainfall) at a daily scale from 2003 to 2017 is used as input data for daily rainfall prediction. As results from statistical performance indicators, it indicates that the SSAGBM model shows a better performance than the EMDGBM model and the original LightGBM algorithm with no decomposition methods. It represents that the accuracy of LightGBM algorithm in rainfall prediction was improved with the SSA method when using multivariate dataset.

• Nuclear Engineering and Technology
• /
• v.56 no.1
• /
• pp.19-33
• /
• 2024

In nuclear thermal-hydraulic system codes, most correlations used for vertical pipes, under downward two-phase flow, have been developed considering small pipes or pool systems. This suggests that there could be uncertainties in applying the correlations to accident scenarios involving large vertical pipes owing to the difference in the characteristics of two-phase flows, or flow conditions, between large and small pipes. In this study, we modified the Multi-dimensional Analysis of Reactor Safety KINS Standard (MARS-KS) code using correlations, such as the drift-flux model and two-phase multiplier, developed in a plant-scale air-inflow experiment conducted for a pipe of diameter 600 mm under downward two-phase flow. The results were then analyzed and compared with those based on previous correlations developed for small pipes and pool conditions. The modified code indicated a good estimation performance in two plant-scale experiments with large pipes. For the siphon-breaking experiment, the maximum errors in water flow for modified and original codes were 2.2% and 30.3%, respectively. For the air-inflow accident experiment, the original code could not predict the trend of frictional pressure gradient in two-phase flow as / increased, while the modified MARS-KS code showed a good estimation performance of the gradient with maximum error of 3.5%.

• Korean Journal of Radiology
• /
• v.23 no.12
• /
• pp.1281-1289
• /
• 2022

Objective: Radiomic modeling using multiple regions of interest in MRI of the brain to diagnose juvenile myoclonic epilepsy (JME) has not yet been investigated. This study aimed to develop and validate radiomics prediction models to distinguish patients with JME from healthy controls (HCs), and to evaluate the feasibility of a radiomics approach using MRI for diagnosing JME. Materials and Methods: A total of 97 JME patients (25.6 ± 8.5 years; female, 45.5%) and 32 HCs (28.9 ± 11.4 years; female, 50.0%) were randomly split (7:3 ratio) into a training (n = 90) and a test set (n = 39) group. Radiomic features were extracted from 22 regions of interest in the brain using the T1-weighted MRI based on clinical evidence. Predictive models were trained using seven modeling methods, including a light gradient boosting machine, support vector classifier, random forest, logistic regression, extreme gradient boosting, gradient boosting machine, and decision tree, with radiomics features in the training set. The performance of the models was validated and compared to the test set. The model with the highest area under the receiver operating curve (AUROC) was chosen, and important features in the model were identified. Results: The seven tested radiomics models, including light gradient boosting machine, support vector classifier, random forest, logistic regression, extreme gradient boosting, gradient boosting machine, and decision tree, showed AUROC values of 0.817, 0.807, 0.783, 0.779, 0.767, 0.762, and 0.672, respectively. The light gradient boosting machine with the highest AUROC, albeit without statistically significant differences from the other models in pairwise comparisons, had accuracy, precision, recall, and F1 scores of 0.795, 0.818, 0.931, and 0.871, respectively. Radiomic features, including the putamen and ventral diencephalon, were ranked as the most important for suggesting JME. Conclusion: Radiomic models using MRI were able to differentiate JME from HCs.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.4
• /
• pp.334-340
• /
• 2002

This paper proposes a self-organizing fuzzy modeling which can create a new hyperplane-shaped cluster by applying multiple regression to input/output data with relatively large fuzzy entropy, add the new cluster to fuzzy rule base and adjust parameters of the fuzzy model in repetition. Tn the coarse tuning, weighted recursive least squared algorithm and fuzzy C-regression model clustering are used and in the fine tuning, gradient descent algorithm is used to adjust parameters of the fuzzy model precisely And learning rates are optimized by utilizing meiosis-genetic algorithm. To check the effectiveness and feasibility of the suggested algorithm, four representative examples for system identification are examined and the performance of the identified fuzzy model is demonstrated in comparison with that of the conventional fuzzy models.

• The Korean Journal of Applied Statistics
• /
• v.19 no.3
• /
• pp.461-479
• /
• 2006

This paper derives joint and conditional Lagrange multiplier tests based on Double-Length Artificial Regression(DLR) for testing functional form and/or the presence of individual(time) effect in a panel regression model. Small sample properties of these tests are assessed by Monte Carlo study, and comparisons are made with LM tests based on Outer Product Gradient(OPG). The results show that the proposed DLR based LM tests have the most appropriate finite sample performance.

• Structural Engineering and Mechanics
• /
• v.45 no.1
• /
• pp.53-67
• /
• 2013

This study aimed to develop a model to accurately predict the acceleration of structural systems during an earthquake. The acceleration and applied force of a structure were measured at current time step and the velocity and displacement were estimated through linear integration. These data were used as input to predict the structural acceleration at next time step. The computation tool used was the Volterra/Wiener neural network (VWNN) which contained the mathematical model to predict the acceleration. For alleviating problems of relatively large-dimensional and nonlinear systems, the VWNN model was utilized as the signal processing tool, including the Taylor series components in the input nodes of the neural network. The number of the intermediate layer nodes in the neural network model, containing the training and simulation stage, was evaluated and optimized. Discussions on the influences of the gradient descent with adaptive learning rate algorithm and the Levenberg-Marquardt algorithm, both for determining the network weights, on prediction errors were provided. During the simulation stage, different earthquake excitations were tested with the optimized settings acquired from the training stage to find out which of the algorithms would result in the smallest error, to determine a proper simulation model.

• International Journal of Computer Science & Network Security
• /
• v.22 no.7
• /
• pp.352-358
• /
• 2022

Text mining is an important process used for analyzing the data collected from different sources like videos, audio, social media, and so on. The tools like Natural Language Processing (NLP) are mostly used in real-time applications. In the earlier research, text mining approaches were implemented using long-short memory (LSTM) networks. In this paper, text mining is performed using average-stochastic gradient descent weight-dropped (AWD)-LSTM techniques to obtain better accuracy and performance. The proposed model is effectively demonstrated by considering the internet movie database (IMDB) reviews. To implement the proposed model Python language was used due to easy adaptability and flexibility while dealing with massive data sets/databases. From the results, it is seen that the proposed LSTM plus weight dropped plus embedding model demonstrated an accuracy of 88.36% as compared to the previous models of AWD LSTM as 85.64. This result proved to be far better when compared with the results obtained by just LSTM model (with 85.16%) accuracy. Finally, the loss function proved to decrease from 0.341 to 0.299 using the proposed model

• Journal of the Korean Society for Geothermal and Hydrothermal Energy
• /
• v.1 no.1
• /
• pp.1-8
• /
• 2005

Aquifer Thermal Energy Storage (ATES) can be a cost-effective and renewable geothermal energy source, depending on site-specific and thermohydraulic conditions. To design an effective ATES system having the effect of groundwater movement, understanding of thermohydraulic processes is necessary. The heat transfer phenomena for an aquifer heat storage are simulated by using FEFLOW with the scenario of heat pump operation with pumping and waste water reinjection in a two layered confined aquifer model. Temperature distribution of the aquifer model is generated, and hydraulic heads and temperature variations are monitored at the both wells during 365 days. The average groundwater velocities are determined with two hydraulic gradient sets according to boundary conditions, and the effect of groundwater flow are shown at the generated thermal distributions of three different depth slices. The generated temperature contour lines at the hydraulic gradient of 0.001 are shaped circular, and the center is moved less than 5 m to the direction of groundwater flow in 365 days simulation period. However at the hydraulic gradient of 0.01, the contour center of the temperature are moved to the end of east boundary at each slice and the largest movement is at bottom slice. By the analysis of thermal interference data between two wells the efficiency of the heat pump system model is validated, and the variation of heads is monitored at injection, pumping and no operation mode.

• IEIE Transactions on Smart Processing and Computing
• /
• v.3 no.5
• /
• pp.259-266
• /
• 2014

In this paper, we propose a novel technique for noise robust automatic speech recognition (ASR). The development of ASR techniques has made it possible to recognize isolated words with a near perfect word recognition rate. However, in a highly noisy environment, a distinct mismatch between the trained speech and the test data results in a significantly degraded word recognition rate (WRA). Unlike conventional ASR systems employing Mel-frequency cepstral coefficients (MFCCs) and a hidden Markov model (HMM), this study employ histogram of oriented gradient (HOG) features and a Support Vector Machine (SVM) to ASR tasks to overcome this problem. Our proposed ASR system is less vulnerable to external interference noise, and achieves a higher WRA compared to a conventional ASR system equipped with MFCCs and an HMM. The performance of our proposed ASR system was evaluated using a phonetically balanced word (PBW) set mixed with artificially added noise.