• Journal of Korea Society of Industrial Information Systems
• /
• v.20 no.3
• /
• pp.61-69
• /
• 2015

This study presents a probabilistic knowledge discovery method to interpret heart rate variability (HRV) based on time and frequency domain indexes, extracted using discrete wavelet transform. The knowledge induction algorithm was composed of two phases: rule generation and rule estimation. Firstly, a rule generation converts numerical attributes to intervals using ROC curve analysis and constructs a reduced ruleset by comparing consistency degree between attribute-value pairs with different decision values. Then, we estimated three measures such as rule support, confidence, and coverage to a probabilistic interpretation for each rule. To show the effectiveness of proposed model, we evaluated the statistical discriminant power of five rules (3 for atrial fibrillation, 1 for normal sinus rhythm, and 1 for both atrial fibrillation and normal sinus rhythm) generated using a data (n=58) collected from 1 channel wireless holter electrocardiogram (ECG), i.e., HeartCall$^{(R)}$, U-Heart Inc. The experimental result showed the performance of approximately 0.93 (93%) in terms of accuracy, sensitivity, specificity, and AUC measures, respectively.

• Korean Journal of Biological Psychiatry
• /
• v.12 no.2
• /
• pp.165-172
• /
• 2005

Objectives:The purpose of this study is to 1) conduct a discrimination analysis of schizophrenia and bipolar affective disorder using MMPI profile through artificial neural network analysis and logistic regression analysis, 2) to make a comparison between advantages and disadvantages of the two methods, and 3) to demonstrate the usefulness of artificial neural network analysis of psychiatric data. Procedure:The MMPI profiles for 181 schizophrenia and bipolar affective disorder patients were selected. Of these profiles, 50 were randomly placed in the learning group and the remaining 131 were placed in the validation group. The artificial neural network was trained using the profiles of the learning group and the 131 profiles of the validation group were analyzed. A logistic regression analysis was then conducted in a similar manner. The results of the two analyses were compared and contrasted using sensitivity, specificity, ROC curves, and kappa index. Results:Logistic regression analysis and artificial neural network analysis both exhibited satisfactory discriminating ability at Kappa index of greater than 0.4. The comparison of the two methods revealed artificial neural network analysis is superior to logistic regression analysis in its discriminating capacity, displaying higher values of Kappa index, specificity, and AUC(Area Under the Curve) of ROC curve than those of logistic regression analysis. Conclusion:Artificial neural network analysis is a new tool whose frequency of use has been increasing for its superiority in nonlinear applications. However, it does possess insufficiencies such as difficulties in understanding the relationship between dependent and independent variables. Nevertheless, when used in conjunction with other analysis tools which supplement it, such as the logistic regression analysis, it may serve as a powerful tool for psychiatric data analysis.

• Journal of Physiology & Pathology in Korean Medicine
• /
• v.32 no.3
• /
• pp.165-170
• /
• 2018

The aim of this study was to investigate the predictive validity of the five organ pattern identification questionnaire(FOPIQ). Data collection was conducted from 190 people who were randomly selected from the general population living in D city from October 2016 to June 2017, and the collected data were analyzed by SPSS 23.0 Statistics Program. Pearson correlation coefficient was used to know the relation between the expert's score and FOPIQ's one. The cut-off value, sensitivity and specificity were analyzed through ROC-curve. Significant p was <.05. The pearson correlation coefficient was .735, .756, .762, .736, and .513 between individual score of FOPIQ and that of the experts in liver, heart, spleen, lung, and kidney, respectively. The cut-off value of the FOPIQ was 46.209, 47.276, 45.336, 48.823, and 42.508 in liver, heart, spleen, lung, and kidney respectively. The AUC derived from the cut-off value of the FOPIQ was .907, .854, .888, .902, and .781 respectively. This study suggests that the FOPIQ could be valid to apply for general population in clinics as well as health checkups.

• Asian Pacific Journal of Cancer Prevention
• /
• v.14 no.7
• /
• pp.4215-4221
• /
• 2013

Purposes: Lung cancer is prevalent worldwide and improvements in timely and effective diagnosis are need. Pentraxin-3 as a novel serum marker for lung cancer (LC) has not been validated in large cohort studies. The aim of the study was to assess its clinical value in diagnosis and prognosis. Methods: We analyzed serum PTX-3 levels in a total of 1,605 patients with LC, benign lung diseases and healthy controls, as well as 493 non-lung cancer patients including 12 different types of cancers. Preoperative and postoperative data were further assessed in patients undergoing LC resection. The diagnostic performance of PTX-3 for LC and early-stage LC was assessed using receiver operating characteristics (ROC) by comparing with serum carcinoembryonic antigen (CEA), cytokeratin 19 fragments (CYFRA 21-1). Results: Levels of PTX-3 in serum were significantly higher in patients with LC than all controls. ROC curves showed the optimum diagnostic cutoff was 8.03ng/mL (AUC 0.823, [95%CI 0.789-0.856], sensitivity 72.8%, and specificity 77.3% in the test cohort; 0.802, [95%CI 0.762-0.843], sensitivity 69.7%, and specificity 76.4% in the validate cohort). Similar diagnostic performance of PTX-3 was observed for early-stage LC. PTX-3 decreased following surgical resection of LC and increased with tumor recurrence. Significantly elevated PTX-3 levels were also seen in patients with non-lung cancers. Conclusions: The present data revealed that PTX-3 was significantly increased in both tissue and serum samples in LC patients. PTX-3 is a valuable biomarker for LC and improved identification of patients with LC and early-stage LC from those with non-malignant lung diseases.

• Asian Pacific Journal of Cancer Prevention
• /
• v.17 no.4
• /
• pp.2177-2183
• /
• 2016

Background: To evaluate the predictive role of a risk of malignancy index in discriminating between benign and malignant adnexal masses preoperatively. Materials and Methods: A total of 408 patients with adnexal masses managed surgically between January 2010 and February 2014 were included. The risk of malignancy indices (RMI) 1, 2, 3 and 4 were calculated using findings for ultrasonography, menopausal status, and CA125 levels. Histopathologic results were the end point. ROC analysis was used for the sensitivity and the specificity of the models. Results: Some 37.6 % of the cases were malignant in the postmenopausal group while 7.9 % were malignant in the premenopausal group. Pelvic pain was the most common complaint, and the majority of the cases were diagnosed at stage 3. The RMI 1, 2, 3 and 4 yielded percentage sensitivities of 76.1, 79.1, 76.1 and 76.1 and specificities of 91.5, 89.1, 90.6, 88.6, respectively. RMI 1 was the most reliable test in the general population according to AUC levels and Kappa statistics. From ROC analysis results of post/premenopausal women, the RMI 1 (cut off: 200) yielded sensitivities of 84.0/60.9 and specificities of 87.7/92.5. With RMI 2 they were 88.6/60.9 and 80.0/91.0, with RMI 3 84.0/60.9 and 87.7/91.8, and with RMI 4 (cut off:400) 81.8/47.8 and 83.6 /44.0. Although test performance of RMI methods were good in a general population and postmenopausal women, the RMI inter-agreement validity was only moderate or fair in premenopausal women. Conclusions: Our study confirms the effectiveness of RMI algorithms in postmenopausal women. However, more sensitive tests are needed for premenopausal women.

• Korean Journal of Community Nutrition
• /
• v.23 no.4
• /
• pp.277-288
• /
• 2018

Objectives: This study was performed to verify the validity and judgment criteria setting of a health status assessment tool based on dietary patterns for middle-aged women. Methods: A total of 474 middle-aged women who visited the Comprehensive Medical Examination Center at Hanmaeum Hospital in Changwon were enrolled (IRB 2013-0005). The validity was verified using clinical indicators for the diagnosis of metabolic syndrome (MS), and it was used to set the criteria for the tool. A logistic regression analysis was performed for validation. The area under-receiver operation (AUC), sensitivity, specificity, and Youden Index were calculated through ROC curve analysis. Statistical analysis was performed by SPSS 21, and p value <0.05 was considered to be statistically significant. Results: The mean score of the group with no MS (73.3 points) was significantly higher compared to the group with MS (65.7 points) (p<0.001). An analysis of the association between the tool scores and risk of MS showed a 0.15-fold reduction in the risk of MS every time the tool's score increased by one point. As the result of the ROC curve analysis, the assessment reference point was set to 71 points, indicating 77.0% sensitivity and 61.0% specificity. Risk of MS was significantly higher in the group with a score of less than 71.0 than a group with more than 71 points (OR=5.28, p<0.001). Conclusions: This study was the first attempt to develop a health status assessment tool based on the dietary patterns for middle-aged women, and this tool has proven its usefulness as an MS assessment tool through the application of middle-aged women in the field of health screening.

• The Korean Journal of Pain
• /
• v.22 no.2
• /
• pp.146-150
• /
• 2009

Background: A skin temperature difference is one of the variables used in the diagnosis of complex regional pain syndrome. However, there have been no reports as to whether the real (${\Delta}T$) or absolute value ($|{\Delta}T|$) of skin temperature differences should be used in the diagnosis of complex regional pain syndrome. This study was conducted to compare the diagnostic validity of ${\Delta}T$ with $|{\Delta}T|$ for complex regional pain syndrome using receiver operating characteristic curves (ROC). Methods: Infrared thermographic images were obtained from the 144 patients who were suspected to have CRPS in a unilateral limb. After ${\Delta}T$ and $|{\Delta}T|$ calculation from the thermographic image, ROCs of ${\Delta}T$ and $|{\Delta}T|$ were developed, and the areas under the curve (AUC) for the ROC curves were compared. Results: AUCs of ${\Delta}T$ and $|{\Delta}T|$ were 0.520 and 0.746 respectively, this difference was statistically significant (P < 0.001). Conclusions: Absolute skin temperature difference shows greater validity in the diagnosis of CRPS than ${\Delta}T$. Therefore, $|{\Delta}T|$ is more useful when comparing the skin temperature of CRPS patients.

• Journal of Convergence for Information Technology
• /
• v.10 no.12
• /
• pp.146-155
• /
• 2020

This study was to evaluate the usefulness of gamma glutamyl transferase (GGT) as a surrogate marker predicting metabolic syndrome. 7,155 non obese men over the age of 20 were studied as subjects. The criteria for diagnosing MetS were the National Cholesterol Education Program - Third Adult Treatment Panel (NCEP-ATP III). The risk of developing MetS according to GGT was conducted logistic regression analysis, and the ROC (receiver operating characteristic) curve was obtained to confirm GGT ability to predict the risk of MetS. Regardless of age and body mass index, MetS had a 7.09 times higher risk of onset in the fourth quartile than in the first quartile of GGT (p<0.001). The AUC (area under the curve) of GGT for the diagnosis of MetS was 0.715, and the cutoff value of GGT was 40.0 U/L, the sensitivity was 65.0%, and the specificity was 70.2%. Therefore, GGT is considered to be a useful diagnostic index for diagnosing MetS.

• The Journal of Korean Medicine
• /
• v.43 no.1
• /
• pp.1-5
• /
• 2022

Objectives: Using algometer, measure the pressure pain threshold (PPT) of the epigastric pain(心下痞硬) and calculate the cut-off value, and this can serve as the basis for prognostic diagnosis of functional dyspepsia so we would like to evaluate its diagnostic value. Methods: We investigated 353 patients with functional dyspepsia symptoms who admitted Gangnam Weedahm Oriental Hospital from February 1, 2021 to February 27, 2021. At the time of the patient's visit, an oriental medical doctor measured the pressure at the first pain point on the Algometer of (CV14), twice each, at 1minute intervals. The ROC (receiver operating characteristic) curve and the optimal cut-off value derived through the diagnosis of the (CV14) PPT value for epigastric pain(心下痞硬) and the gold standard of oriental medical doctor, it was evaluated through. Results: In 353 patients, the area under the ROC curve (AUC) was 0.909 (p=0). In addition, the optimal cutting value was 10.05 (kg/cm2), which was statistically significant. Additionally, the sensitivity of the Algometer's PPT measurement was 0.704 and the specificity was 0.884. As a result, if the PPT value of the Algometer exceeds 10.05 (kg/cm²) in terms of the optimal cutting value, it can be seen that epigastric pain(心下痞硬) is lost. Conclusion: Algometer's PPT value measurement can be a reliable test method for quantification of epigastric pain(心下痞硬) diagnosis and can be useful as an objective indicator.

• Computers and Concrete
• /
• v.32 no.3
• /
• pp.327-337
• /
• 2023

Concrete is the most widely used building material, with various types including high- and ultra-high-strength, reinforced, normal, and lightweight concretes. However, accurately predicting concrete properties is challenging due to the geotechnical design code's requirement for specific characteristics. To overcome this issue, researchers have turned to new technologies like machine learning to develop proper methodologies for concrete specification. In this study, we propose a highly accurate deep learning-based predictive model to investigate the compressive strength (UCS) of lightweight concrete with natural aggregates (pumice). Our model was implemented on a database containing 249 experimental records and revealed that water, cement, water-cement ratio, fine-coarse aggregate, aggregate substitution rate, fine aggregate replacement, and superplasticizer are the most influential covariates on UCS. To validate our model, we trained and tested it on random subsets of the database, and its performance was evaluated using a confusion matrix and receiver operating characteristic (ROC) overall accuracy. The proposed model was compared with widely known machine learning methods such as MLP, SVM, and DT classifiers to assess its capability. In addition, the model was tested on 25 laboratory UCS tests to evaluate its predictability. Our findings showed that the proposed model achieved the highest accuracy (accuracy=0.97, precision=0.97) and the lowest error rate with a high learning rate (R2=0.914), as confirmed by ROC (AUC=0.971), which is higher than other classifiers. Therefore, the proposed method demonstrates a high level of performance and capability for UCS predictions.