Earticle

A model based on genetic algorithm optimization, GA-SVM, is proposed to warn university students of their status. This model improves the predictive effect of support vector machines. The genetic optimization algorithm is used to train the hyperparameters and adjust the kernel parameters, kernel penalty factor C, and gamma to optimize the support vector machine model, which can rapidly achieve convergence to obtain the optimal solution. The experimental model was trained on open-source datasets and validated through comparisons with random forest, backpropagation neural network, and GA-SVM models. The test results show that the genetic algorithm-optimized radial basis kernel support vector machine model GA-SVM can obtain higher accuracy rates when used for early warning in university learning.

The development of data mining technology has pushed data-driven decision-making to gradually become the core content of educational data mining. To identify students who are at risk of failing physical education online courses at an early stage, this article uses bidirectional long-short term memory (BiLSTM) neural networks to construct a deep BiLSTM (DBiLSTM) prediction model. The experimental verification of its effectiveness showed that in the full attribute data experiment, the DBiLSTM specificity at Stage 1 was the highest, at 30.8%, and the accuracy rate at Stage 3 was as high as 73.6%. In the best attribute data experiment, compared to the full attribute, the accuracy of all models at Stage 2 increased, except for the SVM model, which had a 61.8% accuracy rate. At Stage 3, the early warning accuracy of DBiLSTM was higher than other algorithms, with a rate of 75.7%. In the experiment after introducing the balanced data method, the accuracy of the DBiLSTMSMOTE model combined with the Synthetic Minority Oversampling Technique was 72.6%. At this time, the AUC value of DBiLSTM-SMPOTE reached 72.6% in the middle of the semester, significantly superior to other algorithm models. Overall, DBiLSTM is effective in the early warning of students' performance in online sports courses, while DBiLSTM-SMOTE is highly practical in early warning of performance in online sports teaching.

An early warning system was developed to help identify stock status as early as possible. For performance to improve, there needs to be a feature to predict the amount of stock that must be provided and a feature to estimate when to buy goods. This research was conducted to improve the inventory early warning system and optimize the Reminder Block's performance in minimum stock settings. The models used in this study are the single exponential smoothing (SES) method for prediction and the economic order quantity (EOQ) model for determining the quantity. The research was conducted by analyzing the Reminder Block in the early warning system, identifying data needs, and implementing the SES and EOQ mathematical models into the Reminder Block. This research proposes a new Reminder Block that has been added to the SES and EOQ models. It is hoped that this study will help in obtaining accurate information about the time and quantity of repurchases for efficient inventory management.

Enterprises in the coastal regions of China release abundant pollutants that have considerably deteriorated the air quality. To address this issue, an information fusion technology has been proposed herein for predicting coastal air quality in Liaoning Province. To this end, real-time data analysis of water, air, and soil pollutants emitted from diverse coastal enterprises was performed using a multi-model selection strategy for ensemble learning. This approach integrated meteorological information and considered the unique learning principles and observational disparities among various algorithms. The proposed approach explored the influence of collaborative early warning of multi-feature pollution source emissions on the surrounding environment. By combining the base learner and meta-learner in the multi-model fusion strategy, the ensemble model yielded better prediction results, particularly using strong learners at the primary level and linear mode at the secondary level. This optimal combination strategy was used to develop a collaborative monitoring and early warning model, which incorporated multi-feature data from water, air, and soil sources in the coastal environment. This multi-feature collaboration enhanced the prediction accuracy compared with that of single-feature models and further amplified the early warning capabilities enabled by multi-model fusion.

Generators continue to deteriorate in performance due to aging and result in increased failure rates and reduced reliability. Therefore, studies are being conducted on anomaly prediction models for generator engines to prevent potential accidents during operation. However, there are problems in designing the models due to class imbalance and manual input of maintenance history. This study labels data from the time an anomaly occurs up to 60 minutes before the occurrence as anomalies to solve these problems. Data from the time an anomaly occurs up to 30 minutes before the occurrence were also added as derived variables to reflect the warning signs of anomalies in model training. The anomaly prediction models were created using engine log and maintenance history data and applying Random Forest(RF), eXtreme Gradient Boosting(XGB), Linear Support Vector Classifier(LSVC), and Deep Neural Networks(DNN) algorithms. The performance of the models was evaluated by F1-Score and Recall. XGB showed excellent performance in terms of F1-Score, and DNN in terms of Recall. As a result of comparing the F1-Scores to sort the optimal model for each system, XGB was optimal for systems 1, 2, and 4, and RF was optimal for systems 3 and 5. System 5 showed excellent performance when only the derived variable condition was applied, and the other systems showed excellent performance when applying the derived variable and labeling.

본 연구에서는 도시유역에서의 실시간 홍수예경보 목적으로 shot noise process 기반의 강우-유출모형을 제안하였다. 제안된 모형은 각 소유역 별 첨두치, 감쇄상수 및 지체시간으로 결정되는 shot noise의 합으로 표현되며, 기존 강우-유출 모형과는 달리 각 소유역 별 유출량이 독립적으로 유역 출구에 도달하는 구조를 가지고 있다. 제안된 모형의 매개변수는 통상 경험식을 가지고 결정하는 소유역의 집중시간과 저류상수 및 관로에서의 도달시간과 저류상수를 이용하여 쉽게 결정될 수 있는 것으로 확인되었다. 본 연구에서 제안된 모형은 중동 빗물펌프장 배수유역, 구로1 빗물펌프장 배수유역, 대림2 빗물펌프장 배수유역에서 관측된 총 3개의 호우사상에 적용하여 그 성능을 평가하였다. 그 결과를 정리하면 다음과 같다. (1) 본 연구에서 제안된 shot noise process 기반 단위 응답함수는 기존 단위 응답함수와 달리 강우 지속기간에 관계없이 동일한 모양을 갖는다. (2) 제안된 모형의 특성상 강우의 시간간격이 짧을수록 수렴된 결과를 얻을 수 있다. 따라서 도시유역의 특성을 감안할 때 1분이 가장 적절한 것으로 판단된다. (3) Shot noise process 기반 1분 단위 응답함수를 실제 호우사상에 적용하여 유출해석을 수행한 결과, 모의된 유출 수문곡선과 관측 값이 매우 유사한 것으로 나타났다. 이러한 결과는 도시유역에서의 유출해석을 수행하는데 있어 제안된 유출모형이 충분한 적용성이 있다는 것을 보여준다.

This study proposed a rainfall-runoff model for the purpose of real-time flood warning in urban basins. The proposed model was based on the shot noise process, which is expressed as a sum of shot noises determined independently with the peak value, decay parameter and time delay of each sub-basin. The proposed model was different from other rainfall-runoff models from the point that the runoff from each sub-basin reaches the basin outlet independently. The model parameters can be easily determined by the empirical formulas for the concentration time and storage coefficient of a basin and those of the pipe flow. The proposed model was applied to the total of three rainfall events observed at the Jungdong, Guro 1 and Daerim 2 pumping stations to evaluate its applicability. Summarizing the results is as follows. (1) The unit response function of the proposed model, different from other rainfall-runoff models, has the same shape regardless of the rainfall duration. (2) The proposed model shows a convergent shape as the calculation time interval becomes smaller. As the proposed model was proposed to be applied to urban basins, one-minute of calculation time interval would be most appropriate. (3) Application of the one-minute unit response function to the observed rainfall events showed that the simulated runoff hydrographs were very similar to those observed. This result indicates that the proposed model has a good application potential for the rainfall-runoff analysis in urban basins.

교차로에서의 우회전 교통사고가 지속적으로 발생하면서 우회전 교통사고에 대한 대책 마련이 촉구되고 있다. 이에 우회전 지역의 CCTV 영상에서의 객체 탐지를 통해 보행자의 유무를 탐지하고 이를 디스플레이에 경고 문구를 출력해 운전자에게 알리는 기술을 개발하였다. 객체 탐지 모델 중 하나인 YOLO(You Only Look Once) 모델을 이용하여 객체 탐지의 성능평가를 확인하고, 추가적인 후처리 알고리즘을 통해 오인식 문제 해결 및 보행자 확인 시 경고 문구를 출력하는 알고리즘을 개발 하였다. 보행자 혹은 객체를 인식하여 경고 문구를 출력하는 정확 도는 82% 수준으로 측정되었으며 이를 통해 우회전 사고 예방에 기여할 수 있을 것으로 예상된다.

With a continuous occurrence of right-turn traffic accidents at intersections, there is an increasing demand for measures to address these incidents. In response, a technology has been developed to detect the presence of pedestrians through object detection in CCTV footage at right-turn areas and display warning messages on the screen to alert drivers. The YOLO (You Only Look Once) model, a type of object detection model, was employed to assess the performance of object detection. An algorithm was also devised to address misidentification issues and generate warning messages when pedestrians are detected. The accuracy of recognizing pedestrians or objects and outputting warning messages was measured at approximately 82%, suggesting a potential contribution to preventing right-turn accidents

최근 전 세계적으로 기후변화의 영향으로 강우량이 집중되고 강우강도가 커지면서 홍수피해의 규모를 증가시키고 있다. 기존에는 관측되지 않았던 규모의 강우가 내리는가 하면 기록적으로 장기간동안 장마가 지속되기도 한다. 특히, 이러한 피해들은 아세안 국가들에 집중되고 있으며, 최근 해수면 상승, 태풍 및 집중호우로 인해 침수가 빈번히 빌생하는 등 아세안 국가 국민들 중 최소 2,000만 명이 영향을 받고 있다. 우리나라도 각종 ODA사업을 통해 국내의 홍수예경보시스템을 아세안 국가에 지원하고 있지만 통신시설이 불안정하여 중앙제어방식만으로는 한계가 있다. 따라서 본 연구에서는 한 개의 관측소에서 수위, 강우의 관측과, 홍수예측, 경보까지 한번에 가능한 관측소를 개발하기 위한 인공지능기반의 홍수예측모형을 개발하였다. 설마천의 전적비교 관측소의 2009년부터 2020년 까지 10분단위 강우와 수위관측자료를 활용하여 선행예보시간 0.5, 1, 2, 3, 6시간에 대해서 학습, 검증, 시험을 수행하였으며 인공지능알고리즘으로는 LSTM을 적용하였다. 연구결과 모든 선행예보시간에 대해 모형적합도 및 오차에서 우수한 결과를 나타냈다. 설마천과 같이 유역규모가 작고 유역경사가 커서 도달시간이 짧은 경우에는 선행예보시간 1시간은 매우 우수한 예측 결과를 나타낼 것으로 판단되며 유역의 규모나 경사에 따라 더 긴 선행예보시간도 가능할 것으로 예상된다.

In recent years, as rainfall is concentrated and rainfall intensity increases worldwide due to climate change, the scale of flood damage is increasing. Rainfall of a previously unobserved magnitude falls, and the rainy season lasts for a long time on record. In particular, these damages are concentrated in ASEAN countries, and at least 20 million people among ASEAN countries are affected by frequent flooding due to recent sea level rise, typhoons and torrential rain. Korea supports the domestic flood warning system to ASEAN countries through various ODA projects, but the communication network is unstable, so there is a limit to the central control method alone. Therefore, in this study, an artificial intelligence-based flood prediction model was developed to develop an observation station that can observe water level and rainfall, and even predict and warn floods at once at one observation station. Training, validation and testing were carried out for 0.5, 1, 2, 3, and 6 hours of lead time using the rainfall and water level observation data in 10-minute units from 2009 to 2020 at Junjukbi-bridge station of Seolma stream. LSTM was applied to artificial intelligence algorithm. As a result of the study, it showed excellent results in model fit and error for all lead time. In the case of a short arrival time due to a small watershed and a large watershed slope such as Seolma stream, a lead time of 1 hour will show very good prediction results. In addition, it is expected that a longer lead time is possible depending on the size and slope of the watershed.

최근 우리나라에는 기상변화로 인한 영향으로 홍수가 빈번히 발생하고 있다 집중호우는 홍수로 인한 피해를 가중시키는 역할을 하며 매년 많은 사상자와 재산피해를 유발한다. 홍수위험 저감대책에는 구조적인 대책과 비구조적인 대책이 있다. 대부분의 홍수재해 예측의 문제는 비구조적인 대책에 속한다. 신경망 모형은 입력과 출력만을 고려하여 모형을 구성할 수 있기 때문에 비구조적인 문제를 다루기에 적합하다. 인공지능 모형인 신경망 모형을 이용하여 수위예측이 가능한 모형을 구성하고 IHP대표유역 중 하나인 금강 보청천 유역의 기대지점에 적용하였다. 그 결과 신경망 모형은 중소하천유역인 보청천유역에서 홍수위 예측을 위한 우수한 모형으로 판단되었다.

Due to recent unusual climate change, flood happen frequently in Korea. Heavy rainfall increase the damage caused by the flooding. It is cause heavy losses of both life and property every year. Flood hazard mitigation measures consist of structural and non-structural mitigation. Most of flood disaster predictions belong to non-structural mitigation. Neural network is proper to solve non-structural problem. Because it is consider only inputs and outputs to construct model. Real-time water level forecasting model was used to construct artificial intelligence neural network. and it was applied to be a highly suitable tool producing a high water level stage forecasting accuracy at Gidae(No.2) of Bocheong stream, which is IHP representative basins. As a result, neural network was proved to be outstanding model for the water level forecasting in the Bocheong stream catchment.

Circular economy as an advanced development mode for harmonizing the problems among resource, environment and economy, is destined to be the choice for keeping on developing. It is the sustainable development strategy which based on reducing, reusing, recycling principle and the recommendation of circular economy was proposed to build resource-efficient and environmental-friendly society. How to choose the evaluation index system and correctly evaluation the development trance of circular economy is treated as a decision problem. In this paper, a hybrid warning model using matter-element model, combination weight method and place value method for evaluating the development trance of circular economy is proposed. The matter-element model is used to choose the statistical index system of circular economy. The combination weight method is used to calculate the weight of indicators and the place value method is used to contribute the comprehensive evaluation and warning model of circular economy. Finally a case study demonstrates the application of the proposed model.

Circular economy as an advanced development mode for harmonizing the problems among resource, environment and economy, is destined to be the choice for keeping on developing. It is the sustainable development strategy which based on reducing, reusing, recycling principle and the recommendation of circular economy was proposed to build resource-efficient and environmental-friendly society. How to choose the evaluation index system and correctly evaluation the development trance of circular economy is treated as a decision problem. In this paper, a hybrid warning model using matter-element model, combination weight method and place value method for evaluating the development trance of circular economy is proposed. The matter-element model is used to choose the statistical index system of circular economy. The combination weight method is used to calculate the weight of indicators and the place value method is used to contribute the comprehensive evaluation and warning model of circular economy. Finally a case study demonstrates the application of the proposed model.

Along with the development of IoT(Internet of Things) in industrial control field, more and more security issues are emerged, which cause great damage in the physical world. Under the background of IoT for industrial control, we propose a multi-level detection and warning model, the model can find the attacked node quickly and further effectively forecast data transmission situation of IoT. In addition to detecting the attacks accurately and effectively, the algorithm could give different levels of alarm according to network running situations. And then corresponding measures would be taken to guarantee network availability. An OMNeT++( Objective Modular Network Tested in C++) simulation is performed to validate correctness and practicability of the model at last. And the results verify that this model is feasible to a certain degree.

According to the fussy of evaluating index and the uncertainty of evaluating information in flight area, systems thinking tool of complex science management theory - explore graph, is used to construct the fight area risk evaluation index system. And a fight area risk early-warning model based on information entropy attribute reduction and multi-layer BP neural network is proposed. First rough set and information entropy are combined, based on information entropy attribute reduction algorithm, the reduced index information are got. Then based on the multi-layer BP network, the data collected from the flight area was intelligent reasoned and analyzed and evaluated. An example analysis by MATLAB shows that the method is feasible, and it provides support for venture investment project risk management evaluation method.

The health condition of of deep water high pile foundation is vital to the safe operation of bridges. However, pier foundations are vulnerable to damage in deep water due to exposure to sea torrents and corrosive environments over an extended period. In this paper, combined with aninvestigation and analysis of the typical damage characteristics of main pier group pile foundations, we study the safety monitoring and real-time early warning technology of the deep water high pile foundations, we propose an early warning index item and early warning threshold of deep water high pile foundation by utilizing a numerical simulation analysis and referring to domestic and foreign standards and literature. First, we combine the characteristics of structures and draw on more mature evaluation theories and experience in civil engineering-related fields such as dam and bridge engineering. Then, we establish a scheme consisting of a Early Warning Index Systemand evaluation model based on the analytic hierarchy process and constant weight evaluation method and apply the research results to a project based on the Jiashao bridge in Zhejiang province, China. Finally, we verify the rationality and reliability of the Early Warning Index Systemof the Deep Water High Pile Foundations.

The purpose of this study is to develop an Early Warning System (EWS) model to predict a currency crisis in East Asia. The study period extends from 1994 to 2013 in six East Asian countries. In this study, two methods i.e. signal approach as nonparametric model and Logit-Probit as parametric model are used to estimate a currency crisis. It is found that the performance of non-parametric model is adequate to predict the Asian crisis in 1997-1998. Out of 25 macroeconomic indicators observed, 20 indicators are proven to be sensitive to crisis. Meanwhile, during the 2008 global crisis, the non-parametric model’s performance is adequate to predict the crisis in five out of six countries samples. Signal as early warning crisis was issued between 16-21 months prior to the Asian crisis 1997 and 6-9 months prior to the global crisis in 2008. In the parametric EWS model, 11 indicators on logit model and 10 indicators on probit model have correct signs and significant statistics. The empirical result from both EWS model suggests that five macroeconomic indicators namely real effective exchange rate, short-term external debt to reserves, foreign reserves in month of imports, US real interest rate and US growth rate are highly sensitive to the currency crisis in East Asia.

Early warning services for crop diseases are valuable when they provide timely forecasts that farmers can utilize to inform their disease management decisions. In South Korea, collaborative disease controls that utilize unmanned aerial vehicles are commonly performed for most rice paddies. However, such controls could benefit from seasonal disease early warnings with a lead time of a few months. As a first step to establish a seasonal disease early warning service using seasonal climate forecasts, we developed the EPIRICE Daily Risk Model for rice blast by extracting and modifying the core infection algorithms of the EPIRICE model. The daily risk scores generated by the EPIRICE Daily Risk Model were successfully converted into a realistic and measurable disease value through statistical analyses with 13 rice blast incidence datasets, and subsequently validated using the data from another rice blast experiment conducted in Icheon, South Korea, from 1974 to 2000. The sensitivity of the model to air temperature, relative humidity, and precipitation input variables was examined, and the relative humidity resulted in the most sensitive response from the model. Overall, our results indicate that the EPIRICE Daily Risk Model can be used to produce potential disease risk predictions for the seasonal disease early warning service.

Early detection of nuclear proliferation risk is inherently challenging due to the rarity and secrecy of the underlying escalation events. This study presents an investigation of a data-driven early-warning model based on historical data by transforming discrete proliferation stages into a continuous 0-to-1 risk score. The model uses political-economic and HS-code trade variables for 148 countries from 1939 to 2012. The model architecture (i) applies supervised learning regression (LightGBM regressors) in rolling windows to track annual stage scores based on Bleek and Narang taxonomies, (ii) detects residual anomalies using unsupervised learning via Isolation Forests, and (iii) fuses these signals in a meta-classifier to generate interpretable yearly alarm probabilities. The final model achieves an event-F1 score of 0.65 and ROC-AUC of 0.99, with a mean warning lead time of +1.14 years against the historical proliferation events. Quantifying the contributions of inputs to predictions using SHAP analysis reveals a post-1995 shift in proliferation drivers toward uranium centrifuge-related trade patterns. This underscores the need to examine the growing role of global supply chains. The system model based on an explainable, tree-based ensemble machine learning technique provides a transparent, lead-time-positive alternative to "black-box" deep learning and presents the possibility of developing a methodology for integrating multi-modal AI and dynamic data streams into an early warning tool to support nonproliferation intelligence.

최근 5년간 국내 건설현장에서 발생하는 건설 안전사고는 다양한 노력과 투자에도 불구하고 점차 증가하고 있는 추세이다. 이는 건설현장 안전관리에 혁신적인 관리 및 예방 기법의 개발이 시급함을 의미한다. 따라서 본 연구에서는 데이터마이닝 기법 중 하나인 사례기반추론 기법을 이용하여 사전에 발생가능한 건설 안전사고 유형을 예측할 수 있는 안전사고 조기 경보 모델을 제안하였다. 실제 건설안전사고 사례 데이터를 이용하여 모델의 유효성을 검증해 본 결과, 임의로 선택된 테스트데이터에 대하여 매우 유사한 안전사고 유형을 조회하여 제시해주었다. 본 연구의 조기 경보 모델은 건설현장 안전관리자가 발생가능한 안전사고 유형을 예측할 수 있도록 지원하여 건설 안전사고 감소에 기여할 수 있을 것으로 기대된다.

In these days, the accidents on construction site are continuously increasing in spite of the managerial efforts and financial investments. This trend encourage safety managers or decision makers to enhance existing approaches and to develop innovative and proactive management aids. In this study, we propose an early warning model, called the Early Warning Model (hereafter EWM), for the potential types and causes of an accident as a project is progressive, using Case-Based Reasoning technique. The EWM was tested with illustrative cases, and the results showed a few possible cases with high likelihood for a specific situation randomly selected in the construction field. It is expected that the EWM is useful for furnishing safety managers and field supervisors with warning signals during the construction and helpful for assisting them in responding preactively and minimizing the ripple effect of construction site accidents.