• 열처리공학회지
• /
• 제37권3호
• /
• pp.103-113
• /
• 2024

Martensite volume fraction significantly affects the mechanical properties of alloy steels. Martensite start temperature (M_s), transformation temperature for martensite 50 vol.% (M₅₀), and transformation temperature for martensite 90 vol.% (M₉₀) are important transformation temperatures to control the martensite phase fraction. Several researchers proposed empirical equations and machine learning models to predict the Ms temperature. These numerical approaches can easily predict the Ms temperature without additional experiment and cost. However, to control martensite phase fraction more precisely, we need to reduce prediction error of the Ms model and propose prediction models for other martensite transformation temperatures (M₅₀, M₉₀). In the present study, machine learning model was applied to suggest the predictive model for the Ms, M50, M90 temperatures. To explain prediction mechanisms and suggest feature importance on martensite transformation temperature of machine learning models, the explainable artificial intelligence (XAI) is employed. Random forest regression (RFR) showed the best performance for predicting the Ms, M50, M90 temperatures using different machine learning models. The feature importance was proposed and the prediction mechanisms were discussed by XAI.

• 한국인공지능학회지
• /
• 제12권2호
• /
• pp.1-7
• /
• 2024

In this paper, we compare three models (logistic regression, Random Forest, and XGBoost) for predicting stroke occurrence using data from the Korea National Health and Nutrition Examination Survey (KNHANES). We evaluated these models using various metrics, focusing mainly on recall and F1 score to assess their performance. Initially, the logistic regression model showed a satisfactory recall score among the three models; however, it was excluded from further consideration because it did not meet the F1 score threshold, which was set at a minimum of 0.5. The F1 score is crucial as it considers both precision and recall, providing a balanced measure of a model's accuracy. Among the models that met the criteria, XGBoost showed the highest recall rate and showed excellent performance in stroke prediction. In particular, XGBoost shows strong performance not only in recall, but also in F1 score and AUC, so it should be considered the optimal algorithm for predicting stroke occurrence. This study determines that the performance of XGBoost is optimal in the field of stroke prediction.

• Geomechanics and Engineering
• /
• 제37권6호
• /
• pp.539-554
• /
• 2024

This study explores development of prediction model for seismic site classification through the integration of machine learning techniques with horizontal-to-vertical spectral ratio (HVSR) methodologies. To improve model accuracy, the research employs outlier detection methods and, synthetic minority over-sampling technique (SMOTE) for data balance, and evaluates using seven machine learning models using seismic data from KiK-net. Notably, light gradient boosting method (LGBM), gradient boosting, and decision tree models exhibit improved performance when coupled with SMOTE, while Multiple linear regression (MLR) and Support vector machine (SVM) models show reduced efficacy. Outlier detection techniques significantly enhance accuracy, particularly for LGBM, gradient boosting, and voting boosting. The ensemble of LGBM with the isolation forest and SMOTE achieves the highest accuracy of 0.91, with LGBM and local outlier factor yielding the highest F1-score of 0.79. Consistently outperforming other models, LGBM proves most efficient for seismic site classification when supported by appropriate preprocessing procedures. These findings show the significance of outlier detection and data balancing for precise seismic soil classification prediction, offering insights and highlighting the potential of machine learning in optimizing site classification accuracy.

• Journal of Information Technology Applications and Management
• /
• 제31권1호
• /
• pp.1-9
• /
• 2024

Predicting the spread of COVID-19 remains a challenge due to the complexity of the disease and its evolving nature. This study presents an integrated approach using the classic SIR model for infectious diseases, enhanced by the chemical master equation (CME). We employ a Monte Carlo method (SSA) to solve the model, revealing unique aspects of the SARS-CoV-2 virus transmission. The study, a first of its kind in Korea, adopts a step-by-step and complementary approach to model prediction. It starts by analyzing the epidemic's trajectory at local government levels using both basic and stochastic SIR models. These models capture the impact of public health policies on the epidemic's dynamics. Further, the study extends its scope from a single-infected individual model to a more comprehensive model that accounts for multiple infections using the jump SIR prediction model. The practical application of this approach involves applying these layered and complementary SIR models to forecast the course of the COVID-19 epidemic in small to medium-sized local governments, particularly in Gangnam-gu, Seoul. The results from these models are then compared and analyzed.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2021년도 학술발표회
• /
• pp.131-131
• /
• 2021

Deep learning models, especially those based on long short-term memory (LSTM), have presented their superiority in addressing time series data issues recently. This study aims to comprehensively evaluate the performance of deep learning models that belong to the supervised learning category in streamflow prediction. Therefore, six deep learning models-standard LSTM, standard gated recurrent unit (GRU), stacked LSTM, bidirectional LSTM (BiLSTM), feed-forward neural network (FFNN), and convolutional neural network (CNN) models-were of interest in this study. The Red River system, one of the largest river basins in Vietnam, was adopted as a case study. In addition, deep learning models were designed to forecast flowrate for one- and two-day ahead at Son Tay hydrological station on the Red River using a series of observed flowrate data at seven hydrological stations on three major river branches of the Red River system-Thao River, Da River, and Lo River-as the input data for training, validation, and testing. The comparison results have indicated that the four LSTM-based models exhibit significantly better performance and maintain stability than the FFNN and CNN models. Moreover, LSTM-based models may reach impressive predictions even in the presence of upstream reservoirs and dams. In the case of the stacked LSTM and BiLSTM models, the complexity of these models is not accompanied by performance improvement because their respective performance is not higher than the two standard models (LSTM and GRU). As a result, we realized that in the context of hydrological forecasting problems, simple architectural models such as LSTM and GRU (with one hidden layer) are sufficient to produce highly reliable forecasts while minimizing computation time because of the sequential data nature.

• 한국의류학회지
• /
• 제42권3호
• /
• pp.503-515
• /
• 2018

In this research, the three-dimensional structural and colorimetrical modeling of yarn-dyed woven fabrics was conducted based on the Kubelka-Munk theory (K-M theory) for their accurate color predictions. In the K-M theory for textile color formulation, the absorption and scattering coefficients, denoted K and S, respectively, of a colored fabric are represented using those of the individual colorants or color components used. One-hundred forty woven fabric samples were produced in a wide range of structures and colors using red, yellow, green, and blue yarns. Through the optimization of previous two-dimensional color prediction models by considering the key three-dimensional structural parameters of woven fabrics, three three-dimensional K/S-based color prediction models, that is, linear K/S, linear log K/S, and exponential K/S models, were developed. To evaluate the performance of the three-dimensional color prediction models, the color differences, ${\Delta}L^*$, ${\Delta}C^*$, ${\Delta}h^{\circ}$, and ${\Delta}E_{CMC(2:1)}$, between the predicted and the measured colors of the samples were calculated as error values and then compared with those of previous two-dimensional models. As a result, three-dimensional models have proved to be of substantially higher predictive accuracy than two-dimensional models in all lightness, chroma, and hue predictions with much lower ${\Delta}L^*$, ${\Delta}C^*$, ${\Delta}h^{\circ}$, and the resultant ${\Delta}E_{CMC(2:1)}$ values.

• 대한응급의학회지
• /
• 제29권5호
• /
• pp.455-464
• /
• 2018

Objective: Machine learning is not yet widely used in the medical field. Therefore, this study was conducted to compare the performance of preexisting severity prediction models and machine learning based models (random forest [RF], gradient boosting [GB]) for mortality prediction in pneumonia patients. Methods: We retrospectively collected data from patients who visited the emergency department of a tertiary training hospital in Seoul, Korea from January to March of 2015. The Pneumonia Severity Index (PSI) and Sequential Organ Failure Assessment (SOFA) scores were calculated for both groups and the area under the curve (AUC) for mortality prediction was computed. For the RF and GB models, data were divided into a test set and a validation set by the random split method. The training set was learned in RF and GB models and the AUC was obtained from the validation set. The mean AUC was compared with the other two AUCs. Results: Of the 536 investigated patients, 395 were enrolled and 41 of them died. The AUC values of PSI and SOFA scores were 0.799 (0.737-0.862) and 0.865 (0.811-0.918), respectively. The mean AUC values obtained by the RF and GB models were 0.928 (0.899-0.957) and 0.919 (0.886-0.952), respectively. There were significant differences between preexisting severity prediction models and machine learning based models (P<0.001). Conclusion: Classification through machine learning may help predict the mortality of pneumonia patients visiting the emergency department.

• Journal of Korea Trade
• /
• 제25권4호
• /
• pp.17-36
• /
• 2021

Purpose - This study provides useful information to stakeholders by forecasting the tramp shipping market, which is a completely competitive market and has a huge fluctuation in freight rates due to low barriers to entry. Moreover, this study provides the most effective parameters for Baltic Dry Index (BDI) prediction and an optimal model by analyzing and comparing deep learning models such as the artificial neural network (ANN), recurrent neural network (RNN), and long short-term memory (LSTM). Design/methodology - This study uses various data models based on big data. The deep learning models considered are specialized for time series models. This study includes three perspectives to verify useful models in time series data by comparing prediction accuracy according to the selection of external variables and comparison between models. Findings - The BDI research reflecting the latest trends since 2015, using weekly data from 1995 to 2019 (25 years), is employed in this study. Additionally, we tried finding the best combination of BDI forecasts through the input of external factors such as supply, demand, raw materials, and economic aspects. Moreover, the combination of various unpredictable external variables and the fundamentals of supply and demand have sought to increase BDI prediction accuracy. Originality/value - Unlike previous studies, BDI forecasts reflect the latest stabilizing trends since 2015. Additionally, we look at the variation of the model's predictive accuracy according to the input of statistically validated variables. Moreover, we want to find the optimal model that minimizes the error value according to the parameter adjustment in the ANN model. Thus, this study helps future shipping stakeholders make decisions through BDI forecasts.

• 한국해안·해양공학회논문집
• /
• 제34권3호
• /
• pp.72-81
• /
• 2022

본 연구는 제주도 남서부 해역의 표류체 이동 예측을 위해 2020년 8월 제주도 남서부 5개 지점에서 투하된 표층 뜰개 위치자료와 수치모델 예측자료를 학습자료로 이용한 인공지능 기반 입자추적 모델 5개를 구축하였다. 구축된 AI 기법은 기계학습 3종(Extra Trees, LightGBM, Support Vector Machine)과 딥러닝 2종(DNN, RBFN)이다. 또한 해수유동 수치모델 입자추적 예측자료 1종 및 AI 기법 입자추적 예측자료 5종을 표층 뜰개 관측자료와 비교하여 각 예측모델별 예측 정확도를 평가하였다. 6종 모델의 예측 정확도를 평가하기 위해, 5개 정점에 대한 3개 스킬량(MAE, RMSE, NCLS)의 평균값을 비교 검토하였다. 최종적인 결과로서 딥러닝 DNN 모델이 MAE, RMSE, NCLS에서 다른 모델보다 가장 우수하게 나타났다.

• 스마트미디어저널
• /
• 제12권6호
• /
• pp.35-40
• /
• 2023

본 연구는 대화형 추천 시스템인 다중 목표 대화형 추천 시스템(MG-CRS)에서 사용되는 다양한 사전 학습된 언어 모델들을 고찰하고, 각 언어모델의 성능을 비교하고 분석한다. 특히, 언어 모델의 크기가 다중 목표 대화형 추천 시스템의 성능에 어떤 영향을 미치는지에 대해 살펴본다. BERT, GPT2, 그리고 BART의 세 종류의 언어모델을 대상으로 하여, 대표적인 다중 목표 대화형 추천 시스템 데이터셋인 DuRecDial 2.0에서 '타입 예측'과 '토픽 예측'의 정확도를 측정하고 비교한다. 실험 결과, 타입 예측에서는 모든 모델이 뛰어난 성능을 보였지만, 토픽예측에서는 모델 간에 혹은 사이즈에 따라 성능 차이가 관찰되었다. 이러한 결과를 바탕으로 다중 목표 대화형 추천 시스템의 성능 향상을 위한 방향을 제시한다.