• Imaging Science in Dentistry
• /
• v.40 no.2
• /
• pp.63-68
• /
• 2010

Purpose : To compare the CT numbers on 3 cone-beam CT (CBCT) images with those on multi-detector CT (MDCT) image using CT phantom and to develop linear regressive equations using CT numbers to material density for all the CT scanner each. Materials and Methods : Mini CT phantom comprised of five 1 inch thick cylindrical models with 1.125 inches diameter of materials with different densities (polyethylene, polystyrene, plastic water, nylon and acrylic) was used. It was scanned in 3 CBCTs (i-CAT, Alphard VEGA, Implagraphy SC) and 1 MDCT (Somatom Emotion). The images were saved as DICOM format and CT numbers were measured using OnDemand 3D. CT numbers obtained from CBCTs and MDCT images were compared and linear regression analysis was performed for the density, $\rho$ ($g/cm^3$), as the dependent variable in terms of the CT numbers obtained from CBCTs and MDCT images. Results : CT numbers on i-CAT and Implagraphy CBCT images were smaller than those on Somatom Emotion MDCT image (p<0.05). Linear relationship on a range of materials used for this study were $\rho$=0.001H+1.07 with $R^2$ value of 0.999 for Somatom Emotion, $\rho$=0.002H+1.09 with $R^2$ value of 0.991 for Alphard VEGA, $\rho$=0.001H+1.43 with $R^2$ value of 0.980 for i-CAT and $\rho$=0.001H+1.30 with $R^2$ value of 0.975 for Implagraphy. Conclusion: CT numbers on i-CAT and Implagraphy CBCT images were not same as those on Somatom Emotion MDCT image. The linear regressive equations to determine the density from the CT numbers with very high correlation coefficient were obtained on three CBCT and MDCT scan.

• Communications for Statistical Applications and Methods
• /
• v.22 no.2
• /
• pp.137-146
• /
• 2015

Singh (2013) considered the dual problem to the calibration of design weights to obtain a new generalized linear regression estimator (GREG) for the finite population total. In this work, we have made an attempt to suggest a way to use the dual calibration of the design weights in case of multi-auxiliary variables; in other words, we have made an attempt to give an answer to the concern in Remark 2 of Singh (2013) work. The same idea is also used to generalize the GREG estimator proposed by Deville and S$\ddot{a}$rndal (1992). It is not an easy task to find the optimum values of the parameters appear in our approach; therefore, few suggestions are mentioned to select values for such parameters based on a random sample. Based on real data set and under simple random sampling without replacement design, our approach is compared with other approaches mentioned in this paper and for different sample sizes. Simulation results show that all estimators have negligible relative bias, and the multivariate case of Singh (2013) estimator is more efficient than other estimators.

• Geomechanics and Engineering
• /
• v.34 no.5
• /
• pp.507-527
• /
• 2023

Accurately estimation of the geo-mechanical parameters in Artificial Ground Freezing (AGF) is a most important scientific topic in soil improvement and geotechnical engineering. In order for this, one way is using classical and conventional constitutive models based on different theories like critical state theory, Hooke's law, and so on, which are time-consuming, costly, and troublous. The others are the application of artificial intelligence (AI) techniques to predict considered parameters and behaviors accurately. This study presents a comprehensive data-mining-based model for predicting the Young's Modulus of frozen sand under the triaxial test. For this aim, several single and hybrid models were considered including additive regression, bagging, M5-Rules, M5P, random forests (RF), support vector regression (SVR), locally weighted linear (LWL), gaussian process regression (GPR), and multi-layered perceptron neural network (MLP). In the present study, cell pressure, strain rate, temperature, time, and strain were considered as the input variables, where the Young's Modulus was recognized as target. The results showed that all selected single and hybrid predicting models have acceptable agreement with measured experimental results. Especially, hybrid Additive Regression-Gaussian Process Regression and Bagging-Gaussian Process Regression have the best accuracy based on Model performance assessment criteria.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.32 no.2
• /
• pp.93-101
• /
• 2019

The objective of this study is to develop a model which can predict the seismic performance of the slope relatively accurately and efficiently by using artificial neural network(ANN) technique. The quantification of such the seismic performance of the slope is not easy task due to the randomness and the uncertainty of the earthquake input and slope model. Under these circumstances, probabilistic seismic fragility analyses of slope have been carried out by several researchers, and a closed-form equation for slope seismic performance was proposed through a multiple linear regression analysis. However, a traditional statistical linear regression analysis has shown a limit that cannot accurately represent the nonlinearistic relationship between the slope of various conditions and seismic performance. In order to overcome these problems, in this study, we attempted to apply the ANN to generate prediction models of the seismic performance of the slope. The validity of the derived model was verified by comparing this with the conventional multi-linear and multi-nonlinear regression models. As a result, the models obtained through the ANN basically showed excellent performance in predicting the seismic performance of the slope, compared to the models obtained by the statistical regression analyses of the previous study.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.2
• /
• pp.267-275
• /
• 2013

Among data mining techniques, the association rule is the most recently developed technique, and it finds the relevance between two items in a large database. And it is directly applied in the field because it clearly quantifies the relationship between two or more items. When we determine whether an association rule is meaningful, we utilize interestingness measures such as support, confidence, and lift. Interestingness measures are meaningful in that it shows the causes for pruning uninteresting rules statistically or logically. But the criteria of these measures are chosen by experiences, and the number of useful rules is hard to estimate. If too many rules are generated, we cannot effectively extract the useful rules.In this paper, we designed a variety of non-linear regression equations considering all association thresholds between the number of rules and three interestingness measures. And then we diagnosed multi-collinearity and autocorrelation problems, and used analysis of variance results and adjusted coefficients of determination for the best model through numerical experiments.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.27 no.1
• /
• pp.13-21
• /
• 2021

The purpose of this study is to develop the ship valuation model by utilizing the neural network model. The target of the valuation was secondhand VLCC. The variables were set as major factors inducing changes in the value of ship through prior research, and the corresponding data were collected on a monthly basis from January 2000 to August 2020. To determine the stability of subsequent variables, a multi-collinearity test was carried out and finally the research structure was designed by selecting six independent variables and one dependent variable. Based on this structure, a total of nine simulation models were designed using linear regression, neural network regression, and random forest algorithm. In addition, the accuracy of the evaluation results are improved through comparative verification between each model. As a result of the evaluation, it was found that the most accurate when the neural network regression model, which consist of a hidden layer composed of two layers, was simulated through comparison with actual VLCC values. The possible implications of this study first, creative research in terms of applying neural network model to ship valuation; this deviates from the existing formalized evaluation techniques. Second, the objectivity of research results was enhanced from a dynamic perspective by analyzing and predicting the factors of changes in the shipping. market.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.7
• /
• pp.957-963
• /
• 2007

In this paper, Nonlinear Autoregressive (NAR) method based on Least Square-Support Vector Regression (LS-SVR) is introduced and tested for nonlinear sustained vowel modeling. In the database of total 43 sustained vowel of Benign Vocal Fold Lesions having aperiodic waveform, this nonlinear synthesizer near perfectly reproduced chaotic sustained vowels, and also conserved the naturalness of sound such as jitter, compared to Linear Predictive Coding does not keep these naturalness. However, the results of some phonation are quite different from the original sounds. These results are assumed that single-band model can not afford to control and decompose the high frequency components. Therefore multi-band model with wavelet filterbank is adopted for substituting single band model. As a results, multi-band model results in improved stability. Finally, nonlinear sustained vowel modeling using NAR based on LS-SVR can successfully reconstruct synthesized sounds nearly similar to original voiced sounds.

• Journal of Information Processing Systems
• /
• v.17 no.4
• /
• pp.707-720
• /
• 2021

To improve the accuracy of the recommendation system, multi-criteria recommendation systems have been widely researched. However, it is highly complicated to extract the preferred features of users and items from the data. To this end, subjective indicators, which indicate a user's priorities for personalized recommendations, should be derived. In this study, we propose a method for generating recommendation candidates by predicting multi-criteria ratings from reviews and using them to derive user priorities. Using a deep learning model based on convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM), multi-criteria prediction ratings were derived from reviews. These ratings were then aggregated to form a linear regression model to predict the overall rating. This model not only predicts the overall rating but also uses the training weights from the layers of the model as the user's priority. Based on this, a new score matrix for recommendation is derived by calculating the similarity between the user and the item according to the criteria, and an item suitable for the user is proposed. The experiment was conducted by collecting the actual "TripAdvisor" dataset. For performance evaluation, the proposed method was compared with a general recommendation system based on singular value decomposition. The results of the experiments demonstrate the high performance of the proposed method.

• Journal of Korean Society of Water and Wastewater
• /
• v.22 no.2
• /
• pp.213-218
• /
• 2008

Prediction models for post-chlorination require complicated information of reaction time, chlorine dosage considering flow rate as well as environmental conditions such as turbidity, temperature and pH. In order to operate post-chlorination process effectively, the correlations between inlet and outlet of clear well were investigated to develop prediction models of chlorine dosages in post-chlorination process. Correlations of environmental conditions including turbidity and chlorine dosage were investigated to predict residual chlorine at the outlet of clear well. A linear regression model and autoregressive model were developed to apply for the post-chlorination which take place time delay due to detention in clear well tank. The results from autoregressive model show the correlationship of 0.915~0.995. Consequently, the autoregressive model developed in this study would be applicable for real time control for post chlorination process. As a result, the autoregressive model for post chlorination which take place time delay and have multi parameters to control system would contribute to water treatment automation system by applying the process control algorithm.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.65 no.2
• /
• pp.108-113
• /
• 2016

The real-world problems usually show nonlinear and multi-variate characteristics, so it is difficult to establish concrete mathematical models for them. Thus, it is common to practice data-driven modeling techniques in these cases. Among them, most widely adopted techniques are regression model and intelligent model such as neural networks. Regression model has drawback showing lower performance when much non-linearity exists between input and output data. Intelligent model has been shown its superiority to the linear model due to ability capable of effectively estimate desired output in cases of both linear and nonlinear problem. This paper proposes modeling method of daily photovoltaic power systems using ELM(Extreme Learning Machine) based modular networks. The proposed method uses sub-model by fuzzy clustering rather than using a single model. Each sub-model is implemented by ELM. To show the effectiveness of the proposed method, we performed various experiments by dataset acquired during 2014 in real-plant.