• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.703-709
• /
• 2019

This paper proposes speech recognition systems employing Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) structures combined with Hidden Markov Moldel (HMM) to effectively recognize the speech of VeloPharyngeal Insufficiency (VPI) patients, and compares the recognition performance of the systems to the Gaussian Mixture Model (GMM-HMM) and fully-connected Deep Neural Network (DNNHMM) based speech recognition systems. In this paper, the initial model is trained using normal speakers' speech and simulated VPI speech is used for generating a prior model for speaker adaptation. For VPI speaker adaptation, selected layers are trained in the CNN-HMM based model, and dropout regulatory technique is applied in the LSTM-HMM based model, showing 3.68 % improvement in recognition accuracy. The experimental results demonstrate that the proposed LSTM-HMM-based speech recognition system is effective for VPI speech with small-sized speech data, compared to conventional GMM-HMM and fully-connected DNN-HMM system.

• Korean Journal of Remote Sensing
• /
• v.35 no.5_1
• /
• pp.681-692
• /
• 2019

In this paper, a hybrid deep learning model, called 2D convolution with bidirectional long short-term memory (2DCBLSTM), is presented that can effectively combine both spatial and temporal features for crop classification. In the proposed model, 2D convolution operators are first applied to extract spatial features of crops and the extracted spatial features are then used as inputs for a bidirectional LSTM model that can effectively process temporal features. To evaluate the classification performance of the proposed model, a case study of crop classification was carried out using multi-temporal unmanned aerial vehicle images acquired in Anbandegi, Korea. For comparison purposes, we applied conventional deep learning models including two-dimensional convolutional neural network (CNN) using spatial features, LSTM using temporal features, and three-dimensional CNN using spatio-temporal features. Through the impact analysis of hyper-parameters on the classification performance, the use of both spatial and temporal features greatly reduced misclassification patterns of crops and the proposed hybrid model showed the best classification accuracy, compared to the conventional deep learning models that considered either spatial features or temporal features. Therefore, it is expected that the proposed model can be effectively applied to crop classification owing to its ability to consider spatio-temporal features of crops.

• Korean Journal of Construction Engineering and Management
• /
• v.20 no.6
• /
• pp.34-43
• /
• 2019

The maintenance cost for road pavement is gradually increasing due to the continuous increase in road extension as well as increase in the number of old routes that have passed the public period. As a result, there is a need for a method of minimizing costs through preventative grievance preventive maintenance requires the establishment of a strategic plan through accurate prediction of road pavement. Hence, In this study, the deep neural network(DNN) and the recurrent neural network(RNN) were used in order to develop the expressway pavement damage prediction model. A superior model among these two network models was then suggested by comparing and analyzing their performance. In order to solve the RNN's vanishing gradient problem, the LSTM (Long short-term memory) circuits which are a more complicated form of the RNN structure were used. The learning result showed that the RMSE value of the RNN-LSTM model was 0.102 which was lower than the RMSE value of the DNN model, indicating that the performance of the RNN-LSTM model was superior. In addition, high accuracy of the RNN-LSTM model was verified through the comparison between the estimated average road pavement condition and the actually measured road pavement condition of the target section over time.

• Journal of Korea Water Resources Association
• /
• v.54 no.spc1
• /
• pp.1061-1069
• /
• 2021

Recently, the magnitude and frequency of extreme heavy rains and localized heavy rains have increased due to abnormal climate, which caused increased flood damage in river basin. As a result, the nonlinearity of the hydrological system of rivers or basins is increasing, and there is a limitation in that the lead time is insufficient to predict the water level using the existing physical-based hydrological model. This study predicted the water level at Ulsan (Taehwagyo) with a lead time of 0, 1, 2, 3, 6, 12 hours by applying deep learning techniques based on Deep Neural Network (DNN) and Long Short-Term Memory (LSTM) and evaluated the prediction accuracy. As a result, DNN model using the sliding window concept showed the highest accuracy with a correlation coefficient of 0.97 and RMSE of 0.82 m. If deep learning-based water level prediction using a DNN model is performed in the future, high prediction accuracy and sufficient lead time can be secured than water level prediction using existing physical-based hydrological models.

• Journal of Korea Water Resources Association
• /
• v.54 no.9
• /
• pp.667-680
• /
• 2021

We investigated changes in frost days and frost-free periods and to comparatively assess frost event prediction models developed using logistic regression (LR), random forest (RF), and long short-term memory (LSTM) networks. The meteorological variables for the model development were collected from the Suwon, Cheongju, and Gwangju stations for the period of 1973-2019 for spring (March - May) and fall (September - November). The developed models were then evaluated by Precision, Recall, and f-1 score and graphical evaluation methods such as AUC and reliability diagram. The results showed that significant decreases (significance level of 0.01) in the frequencies of frost days were at the three stations in both spring and fall. Overall, the evaluation metrics showed that the performance of RF was highest, while that of LSTM was lowest. Despite higher AUC values (above 0.9) were found at the three stations, reliability diagrams showed inconsistent reliability. A further study is suggested on the improvement of the predictability of both frost events and the first and last frost days by the frost event prediction models and reliability of the models. It would be beneficial to replicate this study at more stations in other regions.

• Korean Chemical Engineering Research
• /
• v.59 no.2
• /
• pp.191-199
• /
• 2021

A distillation column, which is a main facility of the chemical process, separates the desired product from a mixture by using the difference of boiling points. The distillation process requires the optimization and the prediction of operation because it consumes much energy. The target process of this study is difficult to operate efficiently because the composition of feed flow is not steady according to the supplier. To deal with this problem, we could develop a data-driven model to predict operating conditions. However, data preprocessing is essential to improve the predictive performance of the model because the raw data contains outlier and noise. In this study, after optimizing the predictive model based long-short term memory (LSTM) and Random forest (RF), we used a low-pass filter and one-class support vector machine for data preprocessing and compared predictive performance according to the method and range of the preprocessing. The performance of the predictive model and the effect of the preprocessing is compared by using R2 and RMSE. In the case of LSTM, R2 increased from 0.791 to 0.977 by 23.5%, and RMSE decreased from 0.132 to 0.029 by 78.0%. In the case of RF, R2 increased from 0.767 to 0.938 by 22.3%, and RMSE decreased from 0.140 to 0.050 by 64.3%.

• The Journal of the Korea Contents Association
• /
• v.22 no.3
• /
• pp.81-93
• /
• 2022

The stock investing is one of the most popular investment techniques. However, since it is not easy to obtain a return through actual investment, various strategies have been devised and tried in the past to obtain an effective and stable return. Among them, the volatility breakout strategy identifies a strong uptrend that exceeds a certain level on a daily basis as a breakout signal, follows the uptrend, and quickly earns daily returns. It is one of the popular investment strategies that are widely used to realize profits. However, it is difficult to predict stock prices by understanding the price trend pattern of stocks. In this paper, we propose a method of buying and selling stocks by predicting the return in trading based on the volatility breakout strategy using a bi-directional long short-term memory deep neural network that can realize a return in a short period of time. As a result of the experiment assuming actual trading on the test data with the learned model, it can be seen that the results outperform both the return and stability compared to the existing closing price prediction model using the long-short-term memory deep neural network model.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.34 no.4
• /
• pp.128-134
• /
• 2022

Nakdong river estuary is being operated with the goal of expanding the period of seawater inflow from this year to 2022 every month and creating a brackish water area within 15 km of the upstream of the river bank. In this study, the deep learning algorithm Long Short-Term Memory (LSTM) was applied to predict the salinity of the Nakdong Bridge (about 5 km upstream of the river bank) for the purpose of rapid decision making for the target brackish water zone and prevention of salt water damage. Input data were constructed to reflect the temporal and spatial characteristics of the Nakdong River estuary, such as the amount of discharge from Changnyeong and Hamanbo, and an optimal model was constructed in consideration of the hydraulic characteristics of the Nakdong River Estuary by changing the degree according to the sequence length. For prediction accuracy, statistical analysis was performed using the coefficient of determination (R-squred) and RMSE (root mean square error). When the sequence length was 12, the R-squred 0.997 and RMSE 0.122 were the highest, and the prior prediction time showed a high degree of R-squred 0.93 or more until the 12-hour interval.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.28 no.5
• /
• pp.780-790
• /
• 2022

Quantitative risk levels must be presented by analyzing the causes and consequences of accidents and predicting the occurrence patterns of the accidents. For the analysis of marine accidents related to vessel traffic, research on the traffic such as collision risk analysis and navigational path finding has been mainly conducted. The analysis of the occurrence pattern of marine accidents has been presented according to the traditional statistical analysis. This study intends to present a marine accident prediction model using the statistics on marine accidents related to vessel traffic. Statistical data from 1998 to 2021, which can be accumulated by month and hourly data among the Korean domestic marine accidents, were converted into structured time series data. The predictive model was built using a long short-term memory network, which is a representative artificial intelligence model. As a result of verifying the performance of the proposed model through the validation data, the RMSEs were noted to be 52.5471 and 126.5893 in the initial neural network model, and as a result of the updated model with observed datasets, the RMSEs were improved to 31.3680 and 36.3967, respectively. Based on the proposed model, the occurrence pattern of marine accidents could be predicted by learning the features of various marine accidents. In further research, a quantitative presentation of the risk of marine accidents and the development of region-based hazard maps are required.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.5
• /
• pp.1-18
• /
• 2023

With the recent developments in big data and deep learning, a variety of traffic information is collected widely and used for traffic operations. In particular, long short-term memory (LSTM) is used in the field of traffic information prediction with time series characteristics. Since trends, seasons, and cycles differ due to the nature of time series data input for an LSTM, a trial-and-error method based on characteristics of the data is essential for prediction models based on time series data in order to find hyperparameters. If a methodology is established to find suitable hyperparameters, it is possible to reduce the time spent in constructing high-accuracy models. Therefore, in this study, a traffic information prediction model is developed based on highway vehicle detection system (VDS) data and LSTM, and an impact assessment is conducted through changes in the LSTM evaluation indicators for each hyperparameter. In addition, a methodology for finding hyperparameters suitable for predicting highway traffic information in the transportation field is presented.