• Smart Media Journal
• /
• v.7 no.1
• /
• pp.16-23
• /
• 2018

In this paper, we used the Deep Neural Network (DNN) to predict the number of daily spectators of Gwangju - KIA Champions Field in order to provide marketing data for the team and related businesses and for managing the inventories of the facilities in the stadium. In this study, the DNN model, which is based on an artificial neural network (ANN), was used, and four kinds of DNN model were designed along with dropout and batch normalization model to prevent overfitting. Each of four models consists of 10 DNNs, and we added extra models with ensemble model. Each model was evaluated by Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). The learning data from the model randomly selected 80% of the collected data from 2008 to 2017, and the other 20% were used as test data. With the result of 100 data selection, model configuration, and learning and prediction, we concluded that the predictive power of the DNN model with ensemble model is the best, and RMSE and MAPE are 15.17% and 14.34% higher, correspondingly, than the prediction value of the multiple linear regression model.

• Journal of Microbiology and Biotechnology
• /
• v.17 no.9
• /
• pp.1437-1444
• /
• 2007

A response surface model was developed for predicting the growth rates of Staphylococcus aureus in tryptic soy broth (TSB) medium as a function of combined effects of temperature, pH, and NaCl. The TSB containing six different concentrations of NaCl (0, 2, 4, 6, 8, and 10%) was adjusted to an initial of six different pH levels (pH 4, 5, 6, 7, 8, 9, and 10) and incubated at 10, 20, 30, and $40^{\circ}C$. In all experimental variables, the primary growth curves were well ($r^2=0.9000$ to 0.9975) fitted to a Gompertz equation to obtain growth rates. The secondary response surface model for natural logarithm transformations of growth rates as a function of combined effects of temperature, pH, and NaCl was obtained by SAS's general linear analysis. The predicted growth rates of the S. aureus were generally decreased by basic (pH 9-10) or acidic (pH 5-6) conditions and higher NaCl concentrations. The response surface model was identified as an appropriate secondary model for growth rates on the basis of correlation coefficient (r=0.9703), determination coefficient ($r^2=0.9415$), mean square error (MSE=0.0185), bias factor ($B_f=1.0216$), and accuracy factor ($A_f=1.2583$). Therefore, the developed secondary model proved reliable for predictions of the combined effect of temperature, NaCl, and pH on growth rates for S. aureus in TSB medium.

• Journal of Intelligence and Information Systems
• /
• v.23 no.3
• /
• pp.139-153
• /
• 2017

Corporate bankruptcy can cause great losses not only to stakeholders but also to many related sectors in society. Through the economic crises, bankruptcy have increased and bankruptcy prediction models have become more and more important. Therefore, corporate bankruptcy has been regarded as one of the major topics of research in business management. Also, many studies in the industry are in progress and important. Previous studies attempted to utilize various methodologies to improve the bankruptcy prediction accuracy and to resolve the overfitting problem, such as Multivariate Discriminant Analysis (MDA), Generalized Linear Model (GLM). These methods are based on statistics. Recently, researchers have used machine learning methodologies such as Support Vector Machine (SVM), Artificial Neural Network (ANN). Furthermore, fuzzy theory and genetic algorithms were used. Because of this change, many of bankruptcy models are developed. Also, performance has been improved. In general, the company's financial and accounting information will change over time. Likewise, the market situation also changes, so there are many difficulties in predicting bankruptcy only with information at a certain point in time. However, even though traditional research has problems that don't take into account the time effect, dynamic model has not been studied much. When we ignore the time effect, we get the biased results. So the static model may not be suitable for predicting bankruptcy. Thus, using the dynamic model, there is a possibility that bankruptcy prediction model is improved. In this paper, we propose RNN (Recurrent Neural Network) which is one of the deep learning methodologies. The RNN learns time series data and the performance is known to be good. Prior to experiment, we selected non-financial firms listed on the KOSPI, KOSDAQ and KONEX markets from 2010 to 2016 for the estimation of the bankruptcy prediction model and the comparison of forecasting performance. In order to prevent a mistake of predicting bankruptcy by using the financial information already reflected in the deterioration of the financial condition of the company, the financial information was collected with a lag of two years, and the default period was defined from January to December of the year. Then we defined the bankruptcy. The bankruptcy we defined is the abolition of the listing due to sluggish earnings. We confirmed abolition of the list at KIND that is corporate stock information website. Then we selected variables at previous papers. The first set of variables are Z-score variables. These variables have become traditional variables in predicting bankruptcy. The second set of variables are dynamic variable set. Finally we selected 240 normal companies and 226 bankrupt companies at the first variable set. Likewise, we selected 229 normal companies and 226 bankrupt companies at the second variable set. We created a model that reflects dynamic changes in time-series financial data and by comparing the suggested model with the analysis of existing bankruptcy predictive models, we found that the suggested model could help to improve the accuracy of bankruptcy predictions. We used financial data in KIS Value (Financial database) and selected Multivariate Discriminant Analysis (MDA), Generalized Linear Model called logistic regression (GLM), Support Vector Machine (SVM), Artificial Neural Network (ANN) model as benchmark. The result of the experiment proved that RNN's performance was better than comparative model. The accuracy of RNN was high in both sets of variables and the Area Under the Curve (AUC) value was also high. Also when we saw the hit-ratio table, the ratio of RNNs that predicted a poor company to be bankrupt was higher than that of other comparative models. However the limitation of this paper is that an overfitting problem occurs during RNN learning. But we expect to be able to solve the overfitting problem by selecting more learning data and appropriate variables. From these result, it is expected that this research will contribute to the development of a bankruptcy prediction by proposing a new dynamic model.

• Korean Journal of Biological Psychiatry
• /
• v.20 no.2
• /
• pp.54-60
• /
• 2013

Objectives The purpose of this study is to find the influential clinical and physical characteristics which affect apnea-hypopnea index (AHI) in suspected obstructive sleep apnea (OSA) patients. Methods We evaluated the comprehensive factors including sleep related symptoms, clinical scales, medical history, substance use, and anthropometric data of the 119 participants who complained of the symptoms of OSA. All the participants underwent attended-full night laboratory polysomnography. The correlation and multiple regression analysis were conducted to find the influential and predictive factors of AHI. Results A multiple linear regression model 1 showed that higher AHI was associated with higher body mass index (BMI)(p < 0.001) and higher frequency of observed apnea (p = 0.002). In multiple linear regression model 2, AHI was associated with higher BMI (p < 0.001) and loudness of snoring (p = 0.018). Conclusions The present preliminary results suggest that BMI and observed apnea are most influential factors that affect AHI in suspected OSA patients. In the future study we will design the prediction formula for the OSA and AHI, which is useful in the clinical medical field.

• Journal of Korea Water Resources Association
• /
• v.54 no.spc1
• /
• pp.1143-1154
• /
• 2021

The selection of activation function has a great influence on the groundwater level prediction performance of artificial neural network (ANN) model. In this study, five activation functions were applied to ANN model for two groundwater level observation wells in the middle mountainous area of the Pyoseon watershed in Jeju Island. The results of the prediction of the groundwater level were compared and analyzed, and the optimal activation function was derived. In addition, the results of LSTM model, which is a widely used recurrent neural network model, were compared and analyzed with the results of the ANN models with each activation function. As a result, ELU and Leaky ReLU functions were derived as the optimal activation functions for the prediction of the groundwater level for observation well with relatively large fluctuations in groundwater level and for observation well with relatively small fluctuations, respectively. On the other hand, sigmoid function had the lowest predictive performance among the five activation functions for training period, and produced inappropriate results in peak and lowest groundwater level prediction. The ANN-ELU and ANN-Leaky ReLU models showed groundwater level prediction performance comparable to that of the LSTM model, and thus had sufficient potential for application. The methods and results of this study can be usefully used in other studies.

• Journal of Korea Water Resources Association
• /
• v.54 no.spc1
• /
• pp.1095-1105
• /
• 2021

The advancement of dam operation is further required due to the upcoming rainy season, typhoons, or torrential rains. Besides, physical models based on specific rules may sometimes have limitations in controlling the release discharge of dam due to inherent uncertainty and complex factors. This study aims to forecast the water level of the nearest station to the dam multi-timestep-ahead and evaluate the availability when it makes a decision for a release discharge of dam based on LSTM (Long Short-Term Memory) of deep learning. The LSTM model was trained and tested on eight data sets with a 1-hour temporal resolution, including primary data used in the dam operation and downstream water level station data about 13 years (2009~2021). The trained model forecasted the water level time series divided by the six lead times: 1, 3, 6, 9, 12, 18-hours, and compared and analyzed with the observed data. As a result, the prediction results of the 1-hour ahead exhibited the best performance for all cases with an average accuracy of MAE of 0.01m, RMSE of 0.015 m, and NSE of 0.99, respectively. In addition, as the lead time increases, the predictive performance of the model tends to decrease slightly. The model may similarly estimate and reliably predicts the temporal pattern of the observed water level. Thus, it is judged that the LSTM model could produce predictive data by extracting the characteristics of complex hydrological non-linear data and can be used to determine the amount of release discharge from the dam when simulating the operation of the dam.

• Journal of the Korean Geotechnical Society
• /
• v.32 no.7
• /
• pp.35-45
• /
• 2016

Estimation of vertical bearing capacity is critical in the design of bucket foundation used to support offshore structure. Empirical formula and closed form solutions for bucket foundations in uniform sand or clay profiles have been extensively studied. However, the vertical bearing capacity of bucket foundations in alternating layers of sand overlying clay is not well defined. We performed a series of two-dimensional axisymmetric finite element analyses on bucket foundations in sand overlying clay soil, using elasto-plastic soil model. The load transfer mechanism is investigated for various conditions. Performing the parametric study for the friction angles, undrained shear strengths, thickness of sand layer, and aspect ratios of foundation, we present the predictive charts for determining the vertical bearing capacities of bucket foundations in sand overlying clay layer. In addition, after comparing with the finite element analysis results, it is found that linear interpolation between the design charts give acceptable values in these ranges of parameters.

• Journal of the Korean Society of Laryngology, Phoniatrics and Logopedics
• /
• v.10 no.2
• /
• pp.119-123
• /
• 1999

Background and Objectives : Conditions such as muscle atrophy, stretching of strap muscles, and continued craniofacial growth factors have been cited as contributing to the changes observed in the vocal tract structure and function in elderly speakers. The purpose of the present study is to compare F$_1$ and F$_2$ frequency levels in elderly and young adult male and female speakers producing a series of vowels ranging from high-front to low-back placement. Material and Methods : The subjects were two groups of young adults(10 males, 10 females, mean age 21 years old range 19-24 years) and two groups of elderly speakers(10 males, 10 females, mean age 67 years : range 60-84 years). Each subject participated in speech pathologist to be a speaker of unimpared standard Korean. The headphone was positioned 2 cm from the speakers lips. Each speaker sustained the five vowels for 5 s. Formant frequency measures were obtained from an analysis of linear predictive coding in CSL model 4300B(Kay co). Results : Repeated measure AVOVA procedures were completed on the $F_1$ and $F_2$ data for the male and female speakers. $F_2$ formant frequency levels were proven to be significantly lower fir elderly speakers. Conclusions : We presume $F_2$ vocal cavity(from the point of tongue constriction to lip) lengthening in elderly speakers. The research designed to observe dynamic speech production more directly will be needed.

• International Journal of Control, Automation, and Systems
• /
• v.4 no.3
• /
• pp.325-332
• /
• 2006

The dynamic responses of human standing postural control were investigated when subjects were exposed to long-term horizontal vibration. It was hypothesized that the motion of standing posture complexity mainly occurs in the mid-sagittal plane. The motor-driven support platform was designed as a source of vibration. The AC Servo-controlled motors produced anterior/posterior (AP) motion. The platform acceleration and the trunk angular velocity were used as the input and the output of the system, respectively. A method was proposed to identify the complexity of the standing posture dynamics. That is, during AP platform motion, the subject's knee, hip and neck were tightly constrained by fixing assembly, so the lower extremity, trunk and head of the subject's body were individually immovable. Through this method, it was assumed that the ankle joint rotation mainly contributed to maintaining their body balance. Four subjects took part in this study. During the experiment, the random vibration was generated at a magnitude of $0.44m/s^2$, and the duration of each trial was 40 seconds. Measured data were estimated by the coherence function and the frequency response function for analyzing the dynamic behavior of standing control over a frequency range from 0.2 to 3 Hz. Significant coherence values were found above 0.5 Hz. The estimation of frequency response function revealed the dominant resonance frequencies between 0.60 Hz and 0.68 Hz. On the basis of our results illustrated here, the linear model of standing postural control was further concluded.

• Journal of Korean Institute of Industrial Engineers
• /
• v.40 no.1
• /
• pp.8-17
• /
• 2014

In this research, we employed various data mining techniques to build predictive models for win-loss prediction in Korean professional baseball games. The historical data containing information about players and teams was obtained from the official materials that are provided by the KBO website. Using the collected raw data, we additionally prepared two more types of dataset, which are in ratio and binary format respectively. Dividing away-team's records by the records of the corresponding home-team generated the ratio dataset, while the binary dataset was obtained by comparing the record values. We applied seven classification techniques to three (raw, ratio, and binary) datasets. The employed data mining techniques are decision tree, random forest, logistic regression, neural network, support vector machine, linear discriminant analysis, and quadratic discriminant analysis. Among 21(= 3 datasets${\times}$7 techniques) prediction scenarios, the most accurate model was obtained from the random forest technique based on the binary dataset, which prediction accuracy was 84.14%. It was also observed that using the ratio and the binary dataset helped to build better prediction models than using the raw data. From the capability of variable selection in decision tree, random forest, and stepwise logistic regression, we found that annual salary, earned run, strikeout, pitcher's winning percentage, and four balls are important winning factors of a game. This research is distinct from existing studies in that we used three different types of data and various data mining techniques for win-loss prediction in Korean professional baseball games.