• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.569-572
• /
• 2008

The possibility to extent the solution in human detection problem for plug-in on vision-based Human Computer Interaction domain is very attractive, since the successful of the machine leaning theory and computer vision marriage. Bayesian logistic regression is a powerful classifier performing sparseness and high accuracy. The difficulties of finding people in an image will be conquered by implementing this Bavesian model as classifier. The comparison with other massive classifier e.g. SVM and RVM will introduce acceptance of this method for human detection problem. Our experimental results show the good performance of Bavesian logistic regression in human detection problem, both in trade-off curves (ROC, DET) and real-implementation compare to SVM and RVM.

• Korean Journal of Radiology
• /
• v.23 no.6
• /
• pp.664-673
• /
• 2022

Objective: To evaluate the performance of baseline clinical characteristics and pretherapeutic histogram parameters derived from T2 mapping of the extraocular muscles (EOMs) in the prediction of treatment response to intravenous glucocorticoid (IVGC) therapy for active and moderate-to-severe thyroid-associated ophthalmopathy (TAO) and to investigate the effect of fat-suppression (FS) in T2 mapping in this prediction. Materials and Methods: A total of 79 patients clinically diagnosed with active, moderate-to-severe TAO (47 female, 32 male; mean age ± standard deviation, 46.1 ± 10 years), including 43 patients with a total of 86 orbits in the responsive group and 36 patients with a total of 72 orbits in the unresponsive group, were enrolled. Baseline clinical characteristics and pretherapeutic histogram parameters derived from T2 mapping with FS (i.e., FS T2 mapping) or without FS (i.e., conventional T2 mapping) of EOMs were compared between the two groups. Independent predictors of treatment response to IVGC were identified using multivariable analysis. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the predictive performance of the prediction models. Differences between the models were examined using the DeLong test. Results: Compared to the unresponsive group, the responsive group had a shorter disease duration, lower kurtosis (FS-kurtosis), lower standard deviation, larger 75th, 90th, and 95th (FS-95th) T2 relaxation times in FS mapping and lower kurtosis in conventional T2 mapping. Multivariable analysis revealed that disease duration, FS-95th percentile, and FS-kurtosis were independent predictors of treatment response. The combined model, integrating all identified predictors, had an optimized area under the ROC curve of 0.797, 88.4% sensitivity, and 62.5% specificity, which were significantly superior to those of the imaging model (p = 0.013). Conclusion: An integrated combination of disease duration, FS-95th percentile, and FS-kurtosis was a potential predictor of treatment response to IVGC in patients with active and moderate-to-severe TAO. FS T2 mapping was superior to conventional T2 mapping in terms of prediction.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.19
• /
• pp.8447-8450
• /
• 2014

Background: To evaluate factors for predicting the granulosa cell tumor of the ovary (GCTO) pre-operatively. Materials and Methods: This retrospective designed study was conducted on 34 women with GCTO as the study group and 76 women with benign ovarian cysts as the control group. Data were recorded from the hospital database and included age, body mass index (BMI), parity, serum estradiol ($E_2$) levels, diameter of the mass, ultrasonographic features, serum CA125 level, risk of malignancy index (RMI), duration of menopause, postoperative histopathology result, and the neutrophil/lymphocyte ratio (NLR). Results: The demographic parameters showed no statistically significant difference between the groups. Preoperative diameter of the mass, CA125, duration of menopause, and neutrophil/lymphocyte ratio were significantly different between the groups. ROC curve analysis demonstrated that diameter of the mass, serum estradiol and Ca125 levels, RMI and NLR may be discriminative factors in predicting GCTO preoperatively. Conclusions: In conclusion, we think that a careful preoperative workshop including diameter of the mass, serum estradiol ($E_2$) and Ca125 levels, RMI and NLR may predict GCTO and may prevent incomplete approaches.

• International Journal of Advanced Culture Technology
• /
• v.7 no.1
• /
• pp.8-13
• /
• 2019

Despite technical advances in healthcare, the rates of hospital-acquired pressure injury (HAPI) are still high although many are potentially preventable. The purpose of this study was to determine whether tree-based prediction modeling is suitable for assessing the risk of HAPI in ICU patients. Retrospective cohort study has been carried out. A decision tree model was constructed with Age, Weight, eTube, diabetes, Braden score, Isolation, and Number of comorbid conditions as decision nodes. We used RStudio for model training and testing. Correct prediction rate of the final prediction model was 92.4 and the Area Under the ROC curve (AUC) was 0.699, which means there is about 70% chance that the model is able to distinguish between HAPI and non-HAPI. The results of this study has limited generalizability as the data were from a single academic institution. Our research finding shows that the data-driven tree-based prediction modeling may potentially support ICU sensitive risk assessment for HAPI prevention.

• Geomechanics and Engineering
• /
• v.36 no.5
• /
• pp.441-453
• /
• 2024

Sinkhole subsidence and collapse is a common geohazard often formed in karst areas such as the state of Florida, United States of America. To predict the sinkhole occurrence, we need to understand the formation mechanism of sinkhole and its karst hydrogeology. For this purpose, investigating the factors affecting sinkholes is an essential and important step. The main objectives of the presenting study are (1) the development of a machine learning (ML)-based model, namely C5.0 decision tree (C5.0 DT), for the prediction of sinkhole susceptibility, which accounts for sinkhole/subsidence inventory and sinkhole contributing factors (e.g., geological/hydrogeological) and (2) the construction of a regional-scale sinkhole susceptibility map. The study area is east central Florida (ECF) where a cover-collapse type is commonly reported. The C5.0 DT algorithm was used to account for twelve (12) identified hydrogeological factors. In this study, a total of 1,113 sinkholes in ECF were identified and the dataset was then randomly divided into 70% and 30% subsets for training and testing, respectively. The performance of the sinkhole susceptibility model was evaluated using a receiver operating characteristic (ROC) curve, particularly the area under the curve (AUC). The C5.0 model showed a high prediction accuracy of 83.52%. It is concluded that a decision tree is a promising tool and classifier for spatial prediction of karst sinkholes and subsidence in the ECF area.

• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.5
• /
• pp.345-349
• /
• 2009

Social bookmarking systems are a typical web 2.0 service based on folksonomy, providing the platform for storing and sharing bookmarking information. Spammers in social bookmarking systems denote the users who abuse the system for their own interests in an improper way. They can make the entire resources in social bookmarking systems useless by posting lots of wrong information. Hence, it is important to detect spammers as early as possible and protect social bookmarking systems from their attack. In this paper, we applied a diverse set of machine learning approaches, i.e., decision tables, decision trees (ID3), $na{\ddot{i}}ve$ Bayes classifiers, TAN (tree-augment $na{\ddot{i}}ve$ Bayes) classifiers, and artificial neural networks to this task. In our experiments, $na{\ddot{i}}ve$ Bayes classifiers performed significantly better than other methods with respect to the AUC (area under the ROC curve) score as veil as the model building time. Plausible explanations for this result are as follows. First, $na{\ddot{i}}ve$> Bayes classifiers art known to usually perform better than decision trees in terms of the AUC score. Second, the spammer detection problem in our experiments is likely to be linearly separable.

• Korean Journal of Biological Psychiatry
• /
• v.23 no.1
• /
• pp.29-36
• /
• 2016

Objectives Firefighters and rescue workers are likely to be exposed to a variety of traumatic events; as such, they are vulnerable to the risk of post-traumatic stress disorder (PTSD). The psychometric properties of the Korean version of the PTSD Checklist (PCL), a widely used self-report screening tool for PTSD, were assessed in South Korean firefighters and rescue workers. Methods Data were collected via self-report questionnaires and semi-structured clinical interviews administered to 221 firefighters. Internal consistency, item-total correlation, one-week test-retest reliability, convergent validity, and divergent validity were examined. Content validity of the PCL was evaluated using factor analysis and receiver operating characteristic (ROC) analyses were used to estimate the optimal cutoff point and area under the curve. Results The PCL demonstrated excellent internal consistency (${\alpha}=0.97$), item-total correlation (r = 0.72-0.88), test-retest reliability (r = 0.95), and convergent and divergent validity. The total score of PCL was positively correlated with the number of traumatic events experienced (p < 0.001). Factor analysis revealed two theoretically congruent factors: re-experience/avoidance and numbing/hyperarousal. The optimal cutoff was 45 and the area under the ROC curve was 0.97. Conclusions The Korean version of the PCL may be a useful PTSD screening instrument for firefighters and rescue workers, further maximizing opportunities for accurate PTSD diagnosis and treatment.

• The Korean Journal of Pain
• /
• v.36 no.1
• /
• pp.113-127
• /
• 2023

Background: Resting-state functional connectivity (rs-FC) may aid in understanding the link between painmodulating brain regions and the descending pain modulatory system (DPMS) in fibromyalgia (FM). This study investigated whether the differences in rs-FC of the primary somatosensory cortex in responders and non-responders to the conditioned pain modulation test (CPM-test) are related to pain, sleep quality, central sensitization, and the impact of FM on quality of life. Methods: This cross-sectional study included 33 females with FM. rs-FC was assessed by functional magnetic resonance imaging. Change in the numerical pain scale during the CPM-test assessed the DPMS function. Subjects were classified either as non-responders (i.e., DPMS dysfunction, n = 13) or responders (n = 20) to CPM-test. A generalized linear model (GLM) and a receiver operating characteristic (ROC) curve analysis were performed to check the accuracy of the rs-FC to differentiate each group. Results: Non-responders showed a decreased rs-FC between the left somatosensory cortex (S1) and the periaqueductal gray (PAG) (P < 0.001). The GLM analysis revealed that the S1-PAG rs-FC in the left-brain hemisphere was positively correlated with a central sensitization symptom and negatively correlated with sleep quality and pain scores. ROC curve analysis showed that left S1-PAG rs-FC offers a sensitivity and specificity of 85% or higher (area under the curve, 0.78, 95% confidence interval, 0.63-0.94) to discriminate who does/does not respond to the CPM-test. Conclusions: These results support using the rs-FC patterns in the left S1-PAG as a marker for predicting CPM-test response, which may aid in treatment individualization in FM patients.

• Annals of Occupational and Environmental Medicine
• /
• v.35
• /
• pp.24.1-24.15
• /
• 2023

Background: The construction workers are vulnerable to fatigue due to high physical workload. This study aimed to investigate the relationship between overwork and heart rate in construction workers and propose a scheme to prevent overwork in advance. Methods: We measured the heart rates of construction workers at a construction site of a residential and commercial complex in Seoul from August to October 2021 and develop an index that monitors overwork in real-time. A total of 66 Korean workers participated in the study, wearing real-time heart rate monitoring equipment. The relative heart rate (RHR) was calculated using the minimum and maximum heart rates, and the maximum acceptable working time (MAWT) was estimated using RHR to calculate the workload. The overwork index (OI) was defined as the cumulative workload evaluated with the MAWT. An appropriate scenario line (PSL) was set as an index that can be compared to the OI to evaluate the degree of overwork in real-time. The excess overwork index (EOI) was evaluated in real-time during work performance using the difference between the OI and the PSL. The EOI value was used to perform receiver operating characteristic (ROC) curve analysis to find the optimal cut-off value for classification of overwork state. Results: Of the 60 participants analyzed, 28 (46.7%) were classified as the overwork group based on their RHR. ROC curve analysis showed that the EOI was a good predictor of overwork, with an area under the curve of 0.824. The optimal cut-off values ranged from 21.8% to 24.0% depending on the method used to determine the cut-off point. Conclusion: The EOI showed promising results as a predictive tool to assess overwork in real-time using heart rate monitoring and calculation through MAWT. Further research is needed to assess physical workload accurately and determine cut-off values across industries.

• Clinical and Experimental Reproductive Medicine
• /
• v.37 no.2
• /
• pp.115-124
• /
• 2010

Objective: To testify whether the increased peripheral blood natural killer (pbNK) cells fraction and their cytolytic activity could coincide with patient's history of recurrent spontaneous abortion (RSA) and to evaluate these factors are can be valuable diagnostic markers in RSA. Methods: Women with a history of RSA comprised the patient group (n=35). Normal fertile women, who were experienced at least one healthy term birth without history of infertility or recurrent miscarriage, were included as the healthy control group (n=15). The pbNK cells of $CD3^-/CD56^+/CD16^+$ and their cytolytic activities against K562 cells were measured by flow cytometry and the values were compared between study and control groups. Results: Proportions of pbNK cells among peripheral blood monocytes (PBMC) ($14.2{\pm}5.2$ vs. $9.4{\pm}3.7%$, p=0.002, 95% confidence interval [CI], 1.8 to 7.8) was significantly higher in the patient group. The odds ratio of having RSA history was increased as 8.4 folds (59% of sensitivity, 80% of specificity, and 95% CI: 2.0 to 35.8) in patients who showed pbNK cells fraction above 12.1% which was determined as cut-off value by using ROC curve analysis. The cytolytic activities of pbNK cells which measured by three different ratio of effecter pbNK cells to target K562 cells and calculated by the percent of cytolytic K562 cells, were significantly higher in study group than that of control group (in 50:1 ratio, $48.3{\pm}19.0$ vs. $31.3{\pm}11.9%$, p=0.002; in 25:1 ratio, $37.0{\pm}18.1$ vs. $20.2{\pm}9.2%$, p<0.001; in 12.5:1 ratio, $23.5{\pm}12.7$ vs. $12.4{\pm}7.3%$, p=0.001). With the cut-off values of cytolytic activity of pbNK cells as 43.1% (50:1), 26.9% (25:1), and 17.4% (12.5:1) each, the risk of having RSA history was increased by 10.0, 11.4, and 15.0 folds in patients who had increased in each effector of pbNK to target of K562 cells ratio. Conclusion: The analysis of pbNK cells fraction and their cytolytic activity can be valuable diagnostic markers for RSA. We are going to planning the large scaled studies which include the data of obstetric outcomes in subsequent pregnancies to clarify our results of this study.