• 한국학교보건학회지
• /
• 제36권3호
• /
• pp.113-125
• /
• 2023

Purpose: The purpose of this study is to develop a model for predicting mental health risk among adolescents based on health behavior information by employing a hybrid machine learning method. Methods: The study analyzed data of 51,850 domestic middle and high school students from 2022 Youth Health Behavior Survey conducted by the Korea Disease Control and Prevention Agency. Firstly, mental health risk levels (stress perception, suicidal thoughts, suicide attempts, suicide plans, experiences of sadness and despair, loneliness, and generalized anxiety disorder) were classified using the k-mean unsupervised learning technique. Secondly, demographic factors (family economic status, gender, age), academic performance, physical health (body mass index, moderate-intensity exercise, subjective health perception, oral health perception), daily life habits (sleep time, wake-up time, smartphone use time, difficulty recovering from fatigue), eating habits (consumption of high-caffeine drinks, sweet drinks, late-night snacks), violence victimization, and deviance (drinking, smoking experience) data were input to develop a random forest model predicting mental health risk, using logistic and XGBoosting. The model and its prediction performance were compared. Results: First, the subjects were classified into two mental health groups using k-mean unsupervised learning, with the high mental health risk group constituting 26.45% of the total sample (13,712 adolescents). This mental health risk group included most of the adolescents who had made suicide plans (95.1%) or attempted suicide (96.7%). Second, the predictive performance of the random forest model for classifying mental health risk groups significantly outperformed that of the reference model (AUC=.94). Predictors of high importance were 'difficulty recovering from daytime fatigue' and 'subjective health perception'. Conclusion: Based on an understanding of adolescent health behavior information, it is possible to predict the mental health risk levels of adolescents and make interventions in advance.

• 대한기계학회논문집A
• /
• 제26권5호
• /
• pp.969-975
• /
• 2002

Visual effects are important cues for providing occupants with virtual reality in a vehicle simulator which imitates real driving. The viewpoint of an occupant is sensitively dependent upon the occupant's posture, therefore, the total human body motion must be considered in a graphic simulator. A real-time simulation is required for the dynamic analysis of complex human body motion. This study attempts to apply a neural network to the motion analysis in various driving situations. A full car of medium-sized vehicles was selected and modeled, and then analyzed using ADAMS in such driving conditions as bump-pass and lane-change for acquiring the accelerations of chassis of the vehicle model. A hybrid III 50%ile adult male dummy model was selected and modeled in an ellipsoid model. Multibody system analysis software, MADYMO, was used in the motion analysis of an occupant model in the seated position under the acceleration field of the vehicle model. Acceleration data of the head were collected as inputs to the viewpoint movement. Based on these data, a back-propagation neural network was composed to perform the real-time analysis of occupant motions under specified driving conditions and validated output of the composed neural network with MADYMO result in arbitrary driving scenario.

• 한국자동차공학회논문집
• /
• 제8권3호
• /
• pp.139-150
• /
• 2000

Although a number of neck injuries are generated, the data which quantify the kinematic response of the human head and cervical spine in low-speed rear-end automobile collisions is very limited. On this problem, just few in vitro experimental research or some experimental research using dummy on neck injury by rear-end collision was conducted, thus systematic research is requested on full scale injury mechanism. An occupant model for the response of the occupant subject to rear-end collision using commercial dynamics package DADS was developed. Developed model shows more close agreement with the experimental data compared with the MADYMO simulation results for the cases of ${\delta}V=16$ kph in sled test. For the case of ${\delta}V=8$ kph and 33.5 kph with production seat, model also shows its reliable response compared with experimental results using Hybrid III and Hybird III with RID.

• 국제학술발표논문집
• /
• The 6th International Conference on Construction Engineering and Project Management
• /
• pp.674-675
• /
• 2015

A 3D indoor model becomes an indiscernible component of BIM (Building Information Modeling) and GIS (Geographic Information System). However, a huge amount of time and human resources are inevitable for collecting spatial measurements and creating such a 3D indoor model. Also, a varied forms of 3D indoor models exist depending on their purpose of use. Thus, in this study, three different 3D indoor models are defined as 1) omnidirectional images, 2) a 3D realistic model, and 3) 3D indoor as-built model. A series of reconstruction methods is then introduced to construct each type of 3D indoor models: they are an omnidirectional image acquisition method, a hybrid surveying method, and a terrestrial LiDAR-based method. The reconstruction methods are applied to a large and complex atrium, and their 3D modeling results are compared and analyzed.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1995년도 하계학술대회 논문집 B
• /
• pp.992-994
• /
• 1995

FLC has a good performance for complication system or unknown model by using human linguistic method but many part control design are based on expert knowledge or trial-error method and it is difficult to prove stability and robustness of controller. In this paper we improve this problem by setting fuzzy rules by dividing phase plane of error and rate of error change by switching surface. We can guarantee the stability in nonlinear system, and also in fuzzy PID type controller the complexity of controller design is increased by increasing the number of input variables and defining more range of operation if we want performance of more specific rules, thus we need to fine the method to decrease the number of control rules used in FLC design. In this paper the algorithm is validated by simulation using conventional FLC and proposed method.

• Safety and Health at Work
• /
• 제13권3호
• /
• pp.326-335
• /
• 2022

Background: One of the important actions for enhancing human reliability in any industry is assessing human error probability (HEP). The HEART technique is a robust tool for calculating HEP in various industries. The traditional HEART has some weaknesses due to expert judgment. For these reasons, a hybrid model is presented in this study to integrate HEART with Best-Worst Method. Materials Method: In this study, the blasting process in an iron ore mine was investigated as a case study. The proposed HEART-BWM was used to increase the sensitivity of APOA calculation. Then the HEP was calculated using conventional HEART formula. A consistency ratio was calculated using BWM. Finally, for verification of the HEART-BWM, HEP calculation was done by traditional HEART and HEART-BWM. Results: In the view of determined HEPs, the results showed that the mean of HEP in the blasting of the iron ore process was 2.57E-01. Checking the full blast of all the holes after the blasting sub-task was the most dangerous task due to the highest HEP value, and it was found 9.646E-01. On the other side, obtaining a permit to receive and transport materials was the most reliable task, and the HEP was 8.54E-04. Conclusion: The results showed a good consistency for the proposed technique. Comparing the two techniques confirmed that the BWM makes the traditional HEART faster and more reliable by performing the basic comparisons.

• Korean Journal of Radiology
• /
• 제22권12호
• /
• pp.2017-2025
• /
• 2021

Objective: To evaluate the accuracy and clinical efficacy of a hybrid Greulich-Pyle (GP) and modified Tanner-Whitehouse (TW) artificial intelligence (AI) model for bone age assessment. Materials and Methods: A deep learning-based model was trained on an open dataset of multiple ethnicities. A total of 102 hand radiographs (51 male and 51 female; mean age ± standard deviation = 10.95 ± 2.37 years) from a single institution were selected for external validation. Three human experts performed bone age assessments based on the GP atlas to develop a reference standard. Two study radiologists performed bone age assessments with and without AI model assistance in two separate sessions, for which the reading time was recorded. The performance of the AI software was assessed by comparing the mean absolute difference between the AI-calculated bone age and the reference standard. The reading time was compared between reading with and without AI using a paired t test. Furthermore, the reliability between the two study radiologists' bone age assessments was assessed using intraclass correlation coefficients (ICCs), and the results were compared between reading with and without AI. Results: The bone ages assessed by the experts and the AI model were not significantly different (11.39 ± 2.74 years and 11.35 ± 2.76 years, respectively, p = 0.31). The mean absolute difference was 0.39 years (95% confidence interval, 0.33-0.45 years) between the automated AI assessment and the reference standard. The mean reading time of the two study radiologists was reduced from 54.29 to 35.37 seconds with AI model assistance (p < 0.001). The ICC of the two study radiologists slightly increased with AI model assistance (from 0.945 to 0.990). Conclusion: The proposed AI model was accurate for assessing bone age. Furthermore, this model appeared to enhance the clinical efficacy by reducing the reading time and improving the inter-observer reliability.

• Journal of Mechanical Science and Technology
• /
• 제14권3호
• /
• pp.338-349
• /
• 2000

Neck muscle forces and spinal loads at the C4/5 level were estimated that result from isometric voluntary ramp efforts gradually developing to maximums in flexion, extension, left lateral bending and right lateral bending. Electromyographic (EMG) activities, a three-dimensional anatomic data of the neck and a hybrid model, EMG-assisted optimization (EMGAO) model, were used. The model computed the cervical loads at 25%,50%,75%, and 100% of peak moments. The highest model-predicted C4/5 joint compressive forces occurred during flexion; $361\;({\pm}164)\;N,\;811\;({\pm}288)\;N,\;1207\;({\pm}491)\;N\;and\;1674\;({\pm}319)\;N$ in 25%, 50%, 75% and 100% of peak moment respectively. Variations in load distribution among the agonistic muscles and co-contractions of antagonistic muscles were estimated during ramp efforts. Results suggest that higher C4/5 joint loads than previously reported are possible during isometric, voluntary muscle contractions. These higher physiological loads at C4/5 level must be considered possible during orthopedic reconstruction at this level.

• 한국분무공학회지
• /
• 제28권4호
• /
• pp.184-190
• /
• 2023

The European Union has instituted a new emission standard protocol that necessitates real-time measurements from vehicles on actual roads. The adequate development of routes for real driving emissions (RDE) mandates substantial resources, encompassing both vehicles and a portable emission measurement system (PEMS). In this study, a simulation tool was utilized to predict the vehicle speed traversing the routes developed for the RDE measurements. Initially, the vehicle powertrain system was modeled for both a gasoline hybrid vehicle and a gasoline engine-only vehicle. Subsequently, the speed profile for the specified vehicle was constructed based on the RDE route developed for the EURO-6 standard. Finally, the predicted vehicle speed profiles for highway and urban routes were assessed utilizing the actual driving data. The driving model predicted more consistency in the vehicle speed at each driving section. Meanwhile, the human driver tended to accelerate further, and then decelerate in each section, instead of cruising at a predicted section speed.

• 인지과학
• /
• 제12권1_2호
• /
• pp.77-89
• /
• 2001

얼굴 특징점의 지각적 위계구조를 반영한 표정인식 신경망 모형을 설계하였다. 입력자료는 MPEG-4 SNHC(Synthetic/Natural Hybrid Coding)의 얼굴 정의 파라미터(FDP) 중 39개 특징점 각각에 대해 150장의 표정연기 사진을 5개의 크기와 8개의 바위를 갖는 Gabor 필터로분석한 값이었다. 표정영상에 대한 감정상태 평정 값과 39개 특징점의 필터 반응 값을 중가 회귀분석한 결과, 감정상태의 쾌-불쾌 차원은 주로 입과 눈썹 주변의 특징점과 밀접한 과련이 있었고, 각성-수면차원은 주로 눈 주변의 특징점과 밀접한 관련이 있었다. 필터의 크기는 주로 저역 공간 주파수 필터와 감정상태가 관련이 있었고, 필터의 방위는 주로 비스듬한 사선방위와 감정상태가 관련이 있었다. 이를 기초로 표정인식 신경망을 최적화한 결과 원래 1560개(39x5x8) 입력요소를 400개(25x2x8)입력요소로 줄일 수 있었다. 표정인식 신경망의 최적화 결과를 사람의 감정상태 평정과 비교하여 볼 때, 쾌-불쾌 차원에서는 0.886의 상관관계가 있었고, 각성-수면 차원에서는 0.631의 상관관계가 있었다. 표정인식 신경망의 최적화 모형을 기쁨, 슬픔, 놀람, 공포, 분노, 혐오 등의 6가지 기본 정서 범주에 대응한 결과 74%의 인식률을 얻었다. 이러한 결과는 사람의 표정인식 원리를 이용하면 작은 양의 정보로도 최적화된 표정인식 시스템을 구현할수 있다는 점을 시시한다.