• Journal of Intelligence and Information Systems
• /
• v.9 no.1
• /
• pp.227-249
• /
• 2003

Prediction of corporate failure using past financial data is a well-documented topic. Early studies of bankruptcy prediction used statistical techniques such as multiple discriminant analysis, logit and probit. Recently, however, numerous studies have demonstrated that artificial intelligence such as neural networks can be an alternative methodology for classification problems to which traditional statistical methods have long been applied. In building neural network model, the selection of independent and dependent variables should be approached with great care and should be treated as model construction process. Irrespective of the efficiency of a teaming procedure in terms of convergence, generalization and stability, the ultimate performance of the estimator will depend on the relevance of the selected input variables and the quality of the data used. Approaches developed in statistical methods such as correlation analysis and stepwise selection method are often very useful. These methods, however, may not be the optimal ones for the development of neural network model. In this paper, we propose a genetic algorithms approach to find an optimal or near optimal input variables fur neural network modeling. The proposed approach is demonstrated by applications to bankruptcy prediction modeling. Our experimental results show that this approach increases overall classification accuracy rate significantly.

• Earthquakes and Structures
• /
• v.22 no.6
• /
• pp.571-584
• /
• 2022

The residual displacement ratio (RDRs) response spectra have been generally used as an important means to evaluate the post-earthquake repairability, and the ratios of residual to maximum inelastic displacement are considered to be more appropriate for development of the spectra. This methodology, however, assumes that the expected residual displacement can be computed as the product of the RDRs and maximum inelastic displacement, without considering the correlation between these two variables, which inevitably introduces potential systematic error. For providing an adequately accurate estimate of residual displacement, while accounting for the collapse resistance performance prior to the repairability evaluation, a probability-based procedure to estimate the residual displacement demands using the nonlinear static analysis (NSA) is developed for single-degree-of-freedom (SDOF) systems. To this end, the energy-based equivalent damping ratio used for NSA is revised to obtain the maximum displacement coincident with the nonlinear time history analysis (NTHA) results in the mean sense. Then, the possible systematic error resulted from RDRs spectra methodology is examined based on the NTHA results of SDOF systems. Finally, the statistical relation between the residual displacement and the NSA-based maximum displacement is established. The results indicate that the energy-based equivalent damping ratio will underestimate the damping for short period ranges, and overestimate the damping for longer period ranges. The RDRs spectra methodology generally leads to the results being non-conservative, depending on post-yield stiffness. The proposed approach emphasizes that the repairability evaluation should be based on the premise of no collapse, which matches with the current performance-based seismic assessment procedure.

• Journal of Korean Institute of Industrial Engineers
• /
• v.16 no.2
• /
• pp.135-147
• /
• 1990

A nuclear power plant can be viewed as a large complex man-machine system where high system reliability is obtained by ensuring that sub-systems are designed to operate at a very high level of performance. The chance of severe accident involving at least partial core-melt is very low but once it happens the consequence is very catastrophic. The prediction of risk in low probability, high-risk incidents must be examined in the contest of general engineering knowledge and operational experience. Engineering knowledge forms part of the prior information that must be quantified and then updated by statistical evidence gathered from operational experience. Recently, Bayesian procedures have been used to estimate rate of accident and to predict future risks. The Bayesian procedure has advantages in that it efficiently incorporates experts opinions and, if properly applied, it adaptively updates the model parameters such as the rate or probability of accidents. But at the same time it has the disadvantages of computational complexity. The predictive distribution for the time to next incident can not always be expected to end up with a nice closed form even with conjugate priors. Thus we often encounter a numerical integration problem with high dimensions to obtain a predictive distribution, which is practically unsolvable for a model that involves many parameters. In order to circumvent this difficulty, we propose a method of approximation that essentially breaks down a problem involving many integrations into several repetitive steps so that each step involves only a small number of integrations.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.33 no.4
• /
• pp.366-372
• /
• 2009

Creep life prediction has been commonly used by a time-temperature parameter (TTP) which is correlated to an applied stress and temperature, such as Larson-Miller (LM), Orr-Sherby-Dorn (OSD), Manson-Haferd (MH) and Manson-Succop (MS) parameters. A stress-temperature linear model (STLM) based on Arrhenius, Dorn and Monkman-Grant equations was newly proposed through a mathematical procedure. For this model, the logarithm time to rupture was linearly dependent on both an applied stress and temperature. The model parameters were properly determined by using a technique of maximum likelihood estimation of a statistical method, and this model was applied to the creep data of Alloy 617. From the results, it is found that the STLM results showed better agreement than the Eno’s model and the LM parameter ones. Especially, the STLM revealed a good estimation in predicting the long-term creep life of Alloy 617.

• Journal of Power System Engineering
• /
• v.11 no.1
• /
• pp.98-106
• /
• 2007

This paper presents a statistical prediction procedure for the propeller racing of ships with twin screw propellers sailing in ocean waves. The propeller racing is one of the most important factors of seakeeping qualities in relation to the safety of main engine and shafting system. It is especially significant key word for designing the twin-screw-propeller-type ship in view of allowable maximum propeller diameter etc.. In former studies, the propeller racing generally means the situation (propeller exposed) in which the relative motion amplitude between ship hull and wave surface would exceed a depth of point in rotary disk propeller. Therefore, it seems that the magnitude of the amplitude and its exceeding frequency have been examined as a principal subject of study as usual. However, the time during which the amplitude exceeds a depth of point must be also one of most important factor affecting the trend of propeller racing. This paper proposes a simply practical method for estimating the time lasting of exposed propeller related to twin screw propeller racing in rough confused seas on the basis of the statistics. Then, it is confirmed that the practical method is useful and convenience for considering the propeller racing in the stage of the basic design.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2003.05a
• /
• pp.51-59
• /
• 2003

Marine Casualty Forecasting System (MCFS) is to broadcast the prediction number and risk level of marine casualties as like daily weather forecasting. The MCFS consists of marine casualty numerical D/B, prediction model and, three-dimensional statistics visualization system. The implementation procedure for the numerical D/B is described in the paper. The data relating to a total of 724 ship casualties in the west-southern sea area (latitude 33$^{\circ}$N∼35$^{\circ}$ and longitude 124$^{\circ}$E∼127$^{\circ}$E) of Korean peninsula for 11 years (1999∼2000) have been compiled. The analysis method of numerical D/B is proposed and discussed its usability.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2000.11a
• /
• pp.307-310
• /
• 2000

Recently, the demand for the Floating, Production Storage, and Offloading facility(FPSO) which has some economic and technical advantages, has increased in offshore oil production areas. The basic characteristics of a 343,000 DWT class FPSO which is being built in Hyundai Heavy Industries and shall be installed in offshore Angola, is almost same as that of oil carriers. However, she do not have self-propulsion system, but has additional facilities for oil production and positioning system. Main noise source contributing to the cabin noise of the accommodation, are classified into the machine in the engine room and the deckhouse, HVAC system, and the topside equipments. In general, the noise regulation for the offshore structure is much severer than that of the common commercial ships and the maximum acceptable sound pressure level of cabins is specified in 45dB(A). This paper describes the procedure of noise analysis along with its results. Noise analysis has been carried out for the case of emergency diesel generator running condition and the case of normal production condition and the results has been compared with the measurement results of the first case. Based on the results, proper countermeasures to reduce excessive noise level has been applied considering the characteristics of sources and receiver spaces and can be satisfied the specifications at all spaces.

• Earthquakes and Structures
• /
• v.25 no.6
• /
• pp.469-479
• /
• 2023

A probabilistic seismic damage analysis is an essential procedure to identify seismically vulnerable structures, prioritize the seismic retrofit, and ultimately minimize the overall seismic risk. To assess the seismic risk of multiple structures within a region, a large number of nonlinear time-history structural analyses must be conducted and studied. As a result, each assessment requires high computing resources. To overcome this limitation, we explore a deep learning-based metamodel to enable the prediction of the mean and the standard deviation of the seismic damage distribution of track-on steel-plate girder railway bridges in Korea considering the geometric variation. For machine learning training, nonlinear dynamic time-history analyses are performed to generate 800 high-fidelity datasets on the seismic response. Through intensive trial and error, the study is concentrated on developing an optimal machine learning architecture with the pre-identified variables of the physical configuration of the bridge. Additionally, the prediction performance of the proposed method is compared with a previous, well-defined, response surface model. Finally, the statistical testing results indicate that the overall performance of the deep-learning model is improved compared to the response surface model, as its errors are reduced by as much as 61%. In conclusion, the model proposed in this study can be effectively deployed for the seismic fragility and risk assessment of a region with a large number of structures.

• 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 the Semiconductor & Display Technology
• /
• v.18 no.4
• /
• pp.30-34
• /
• 2019

We analyzed how the features in plasma information based virtual metrology (PI-VM) for SiO₂ etching depth with variation of 5% contribute to the prediction accuracy, which is previously developed by Jang. As a single feature, the explanatory power to the process results is in the order of plasma information about electron energy distribution function (PI_EEDF), equipment, and optical emission spectroscopy (OES) features. In the procedure of stepwise variable selection (SVS), OES features are selected after PI_EEDF. Informative vector for developed PI-VM also shows relatively high correlation between OES features and etching depth. This is because the reaction rate of each chemical species that governs the etching depth can be sensitively monitored when OES features are used with PI_EEDF. Securing PI_EEDF is important for the development of virtual metrology (VM) for prediction of process results. The role of PI_EEDF as an independent feature and the ability to monitor variation of plasma thermal state can make other features in the procedure of SVS more sensitive to the process results. It is expected that fault detection and classification (FDC) can be effectively developed by using the PI-VM.