• Journal of Intelligence and Information Systems
• /
• v.29 no.3
• /
• pp.419-437
• /
• 2023

As the number of thin filers in Korea surpasses 12 million, there is a growing interest in enhancing the accuracy of assessing their credit default risk to generate additional revenue. Specifically, researchers are actively pursuing the development of default prediction models using machine learning and deep learning algorithms, in contrast to traditional statistical default prediction methods, which struggle to capture nonlinearity. Among these efforts, Graph Neural Network (GNN) architecture is noteworthy for predicting default in situations with limited data on thin filers. This is due to their ability to incorporate network information between borrowers alongside conventional credit-related data. However, prior research employing graph neural networks has faced limitations in effectively handling diverse categorical variables present in credit information. In this study, we introduce the Transformer embedded Graph Convolutional Network (TeGCN), which aims to address these limitations and enable effective default prediction for thin filers. TeGCN combines the TabTransformer, capable of extracting contextual information from categorical variables, with the Graph Convolutional Network, which captures network information between borrowers. Our TeGCN model surpasses the baseline model's performance across both the general borrower dataset and the thin filer dataset. Specially, our model performs outstanding results in thin filer default prediction. This study achieves high default prediction accuracy by a model structure tailored to characteristics of credit information containing numerous categorical variables, especially in the context of thin filers with limited data. Our study can contribute to resolving the financial exclusion issues faced by thin filers and facilitate additional revenue within the financial industry.

• Journal of Service Research and Studies
• /
• v.13 no.2
• /
• pp.99-113
• /
• 2023

In a customer-based industry, customer retention is the competitiveness of a company, and improving customer retention improves the competitiveness of the company. Therefore, accurate prediction and management of potential dormant customers is paramount to increasing the competitiveness of the enterprise. In particular, there are numerous competitors in the domestic card industry, and the government is introducing an automatic closing system for dormant card management. As a result of these social changes, the card industry must focus on better predicting and managing potential dormant cards, and better predicting dormant customers is emerging as an important challenge. In this study, the Recurrent Neural Network (RNN) methodology was used to predict potential dormant customers in the card industry, and in particular, Long-Short Term Memory (LSTM) was used to efficiently learn data for a long time. In addition, to redefine the variables needed to predict dormant customers in the card industry, Unified Theory of Technology (UTAUT), an integrated technology acceptance theory, was applied to redefine and group the variables used in the model. As a result, stable model accuracy and F-1 score were obtained, and Hit-Ratio proved that models using LSTM can produce stable results compared to other algorithms. It was also found that there was no moderating effect of demographic information that could occur in UTAUT, which was pointed out in previous studies. Therefore, among variable selection models using UTAUT, dormant customer prediction models using LSTM are proven to have non-biased stable results. This study revealed that there may be academic contributions to the prediction of dormant customers using LSTM algorithms that can learn well from previously untried time series data. In addition, it is a good example to show that it is possible to respond to customers who are preemptively dormant in terms of customer management because it is predicted at a time difference with the actual dormant capture, and it is expected to contribute greatly to the industry.

• Journal of Korea Water Resources Association
• /
• v.56 no.4
• /
• pp.245-259
• /
• 2023

The agricultural reservoir is a critical water supply system in South Korea, providing approximately 60% of the agricultural water demand. However, the reservoir faces several issues that jeopardize its efficient operation and management. To address this issues, we propose a novel deep-learning-based water level recognition model that uses CCTV image data to accurately estimate water levels in agricultural reservoirs. The model consists of three main parts: (1) dataset construction, (2) image segmentation using the U-Net algorithm, and (3) CCTV-based water level recognition using either CNN or ResNet. The model has been applied to two reservoirs G-reservoir and M-reservoir with observed CCTV image and water level time series data. The results show that the performance of the image segmentation model is superior, while the performance of the water level recognition model varies from 50 to 80% depending on water level classification criteria (i.e., classification guideline) and complexity of image data (i.e., variability of the image pixels). The performance of the model can be improved if more numbers of data can be collected.

• Journal of Intelligence and Information Systems
• /
• v.29 no.2
• /
• pp.35-56
• /
• 2023

With the development of information and communication technology, numerous reviews are continuously posted on websites, which causes information overload problems. Therefore, users face difficulty in exploring reviews for their decision-making. To solve such a problem, many studies on review helpfulness prediction have been actively conducted to provide users with helpful and reliable reviews. Existing studies predict review helpfulness mainly based on the features included in the review. However, such studies disable providing the reason why predicted reviews are helpful. Therefore, this study aims to propose a methodology for applying eXplainable Artificial Intelligence (XAI) techniques in review helpfulness prediction to address such a limitation. This study uses restaurant reviews collected from Yelp.com to compare the prediction performance of six models widely used in previous studies. Next, we propose an explainable review helpfulness prediction model by applying the XAI technique to the model with the best prediction performance. Therefore, the methodology proposed in this study can recommend helpful reviews in the user's purchasing decision-making process and provide the interpretation of why such predicted reviews are helpful.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.21 no.1
• /
• pp.105-122
• /
• 2022

According to the statistics about the fatal crashes that have occurred on the expressways for the last 5 years, those who died on the shoulders of the road has been as 3 times high as the others who died on the expressways. It suggests that the crashes on the shoulders of the road should be fatal, and that it would be important to prevent the traffic crashes by cracking down on the vehicles intruding the shoulders of the road. Therefore, this study proposed a method to detect a vehicle that violates the shoulder lane by using the Faster R-CNN. The vehicle was detected based on the Faster R-CNN, and an additional reading module was configured to determine whether there was a shoulder violation. For experiments and evaluations, GTAV, a simulation game that can reproduce situations similar to the real world, was used. 1,800 images of training data and 800 evaluation data were processed and generated, and the performance according to the change of the threshold value was measured in ZFNet and VGG16. As a result, the detection rate of ZFNet was 99.2% based on Threshold 0.8 and VGG16 93.9% based on Threshold 0.7, and the average detection speed for each model was 0.0468 seconds for ZFNet and 0.16 seconds for VGG16, so the detection rate of ZFNet was about 7% higher. The speed was also confirmed to be about 3.4 times faster. These results show that even in a relatively uncomplicated network, it is possible to detect a vehicle that violates the shoulder lane at a high speed without pre-processing the input image. It suggests that this algorithm can be used to detect violations of designated lanes if sufficient training datasets based on actual video data are obtained.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.6
• /
• pp.260-268
• /
• 2024

Cervical cancer is the fourth most common cancer in women worldwide, and more than 604,000 new cases were reported in 2020 alone, resulting in approximately 341,831 deaths. The Cox regression model is a major model widely adopted in cancer research, but considering the existence of nonlinear associations, it faces limitations due to linear assumptions. To address this problem, this paper proposes ResSurvNet, a new model that improves the accuracy of cervical cancer mortality prediction using ResNet's residual learning framework. This model showed accuracy that outperforms the DNN, CPH, CoxLasso, Cox Gradient Boost, and RSF models compared in this study. As this model showed accuracy that outperformed the DNN, CPH, CoxLasso, Cox Gradient Boost, and RSF models compared in this study, this excellent predictive performance demonstrates great value in early diagnosis and treatment strategy establishment in the management of cervical cancer patients and represents significant progress in the field of survival analysis.

• Information Systems Review
• /
• v.25 no.3
• /
• pp.59-81
• /
• 2023

Elderly suicide problem has become worse in South Korea. With a rapid aging of the population, the trend of suicide among the elderly is expected to accelerate, preventing elderly suicide has been considered an important societal problem. Thus, we aim to investigate various factors that explain suicidal ideation and to develop a predictive model for suicidal ideation in the context of elderly people in South Korea. To this end, this study contributes to addressing the elderly suicide problem. By using seven-year panel data from the Korea Welfare Panel Survey, we extract various potential causal factors for elderly suicidal ideation based on interpersonal theory of suicide and social disorganization theory. Then a panel logit model was employed to assess the impacts of potential factors on suicidal ideation and deep learning and machine learning algorithms were used to develop a predictive model for suicidal ideation of elderly people. The results of our study provide practical implications for preventing elderly suicide by identifying causal factors of suicidal ideation and a high suicidal risk group of the elderly. This study sheds light on synergy of mixed methodology and provides various academic implications.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.7
• /
• pp.291-298
• /
• 2024

Despite its wide application, cloud computing raises privacy leakage concerns because users should send their private data to the cloud. Homomorphic encryption (HE) can resolve the concerns by allowing cloud servers to compute on encrypted data without decryption. However, due to the huge computation overhead of HE, simply executing an entire cloud program with HE causes significant computation. Manually partitioning the program and applying HE only to the partitioned program for the cloud can reduce the computation overhead. However, the manual code partitioning and HE-transformation are time-consuming and error-prone. This work proposes a new homomorphic encryption enabled annotation-guided code partitioning compiler, called Heapa, for privacy preserving cloud computing. Heapa allows programmers to annotate a program about the code region for cloud computing. Then, Heapa analyzes the annotated program, makes a partition plan with a variable list that requires communication and encryption, and generates a homomorphic encryptionenabled partitioned programs. Moreover, Heapa provides not only two region-level partitioning annotations, but also two instruction-level annotations, thus enabling a fine-grained partitioning and achieving better performance. For six machine learning and deep learning applications, Heapa achieves a 3.61 times geomean performance speedup compared to the non-partitioned cloud computing scheme.

• Phonetics and Speech Sciences
• /
• v.16 no.3
• /
• pp.79-86
• /
• 2024

Transformer models have shown remarkable performance in extracting meaningful information from sequential input data such as text and images, and are gaining attention as end-to-end models for speech recognition. This study compared the performances of the Transformer speech recognition model and its enhanced versions, the Conformer and E-Branchformer, when applied to Korean speech recognition. Using Korean speech data from AIHub, we prepared a training set of approximately 7,500 hours and evaluated the models using the ESPnet toolkit. Additionally, we compared syllables and subwords as recognition units and analyzed the performance differences with changes in the number of tokens using Byte Pair Encoding. The results showed that the E-Branchformer achieved the best performance in Korean speech recognition and Conformer outperformed Transformer but degraded in performance for long utterances owing to cross-attention alignment errors. We aimed to determine the optimal settings by analyzing the performance changes with subword token adjustments. This study comprehensively evaluated model accuracy and processing speed to maximize the efficiency of Korean speech recognition. This is expected to contribute to the training of large-scale Korean speech recognition models and improve Conformer recognition errors. Future research should include additional experiments with diverse Korean speech datasets and enhance the recognition performance through structural improvements in the Conformer.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.40 no.3
• /
• pp.273-283
• /
• 2020

Because of climate change, the occurrence of localized and heavy rainfall is increasing. It is important to predict floods in urban areas that have suffered inundation in the past. For flood prediction, not only numerical analysis models but also machine learning-based models can be applied. The LSTM (Long Short-Term Memory) neural network used in this study is appropriate for sequence data, but it demands a lot of data. However, rainfall that causes flooding does not appear every year in a single urban basin, meaning it is difficult to collect enough data for deep learning. Therefore, in addition to the rainfall observed in the study area, the observed rainfall in another urban basin was applied in the predictive model. The LSTM neural network was used for predicting the total overflow, and the result of the SWMM (Storm Water Management Model) was applied as target data. The prediction of the inundation map was performed by using logistic regression; the independent variable was the total overflow and the dependent variable was the presence or absence of flooding in each grid. The dependent variable of logistic regression was collected through the simulation results of a two-dimensional flood model. The input data of the two-dimensional flood model were the overflow at each manhole calculated by the SWMM. According to the LSTM neural network parameters, the prediction results of total overflow were compared. Four predictive models were used in this study depending on the parameter of the LSTM. The average RMSE (Root Mean Square Error) for verification and testing was 1.4279 ㎥/s, 1.0079 ㎥/s for the four LSTM models. The minimum RMSE of the verification and testing was calculated as 1.1655 ㎥/s and 0.8797 ㎥/s. It was confirmed that the total overflow can be predicted similarly to the SWMM simulation results. The prediction of inundation extent was performed by linking the logistic regression with the results of the LSTM neural network, and the maximum area fitness was 97.33 % when more than 0.5 m depth was considered. The methodology presented in this study would be helpful in improving urban flood response based on deep learning methodology.