• Geomechanics and Engineering
• /
• v.25 no.1
• /
• pp.1-16
• /
• 2021

Machine learning models have been widely used for landslide susceptibility assessment (LSA) in recent years. The large number of inputs or conditioning factors for these models, however, can reduce the computation efficiency and increase the difficulty in collecting data. Feature selection is a good tool to address this problem by selecting the most important features among all factors to reduce the size of the input variables. However, two important questions need to be solved: (1) how do feature selection methods affect the performance of machine learning models? and (2) which feature selection method is the most suitable for a given machine learning model? This paper aims to address these two questions by comparing the predictive performance of 13 feature selection-based machine learning (FS-ML) models and 5 ordinary machine learning models on LSA. First, five commonly used machine learning models (i.e., logistic regression, support vector machine, artificial neural network, Gaussian process and random forest) and six typical feature selection methods in the literature are adopted to constitute the proposed models. Then, fifteen conditioning factors are chosen as input variables and 1,017 landslides are used as recorded data. Next, feature selection methods are used to obtain the importance of the conditioning factors to create feature subsets, based on which 13 FS-ML models are constructed. For each of the machine learning models, a best optimized FS-ML model is selected according to the area under curve value. Finally, five optimal FS-ML models are obtained and applied to the LSA of the studied area. The predictive abilities of the FS-ML models on LSA are verified and compared through the receive operating characteristic curve and statistical indicators such as sensitivity, specificity and accuracy. The results showed that different feature selection methods have different effects on the performance of LSA machine learning models. FS-ML models generally outperform the ordinary machine learning models. The best FS-ML model is the recursive feature elimination (RFE) optimized RF, and RFE is an optimal method for feature selection.

• Journal of the Korea Convergence Society
• /
• v.10 no.5
• /
• pp.125-131
• /
• 2019

This study tries to grasp the factor that affects the acceptance of beacon as an O2O marketing tool. This study examined whether there is a difference between beacon accepter as a means of marketing communication and non-accepter in terms of related variables. As a result, there were significant differences between beacon perceiver and non-perceiver in 'smartphone usage' and 'brand consciousness'. In order to understand the predictive variables influencing beacon acceptance as a means of marketing communication, this study used 'perceiving beacon as a marketing communication media' as a dependent variable to perform logistic regression analysis. As a result of the analysis, it was found that 'smartphone usage' and 'brand consciousness' were the predictive variables affecting beacon perceiving. This study tried to analysis the results in the viewpoint of perceived usefulness and ease-of-use which were insisted by TAM.

• Computers and Concrete
• /
• v.27 no.4
• /
• pp.319-332
• /
• 2021

The performance of gene expression programming (GEP) in predicting the compressive strength of bacteria-incorporated geopolymer concrete (GPC) was examined in this study. Ground-granulated blast-furnace slag (GGBS), new bacterial strains, fly ash (FA), silica fume (SF), metakaolin (MK), and manufactured sand were used as ingredients in the concrete mixture. For the geopolymer preparation, an 8 M sodium hydroxide (NaOH) solution was used, and the ambient curing temperature (28℃) was maintained for all mixtures. The ratio of sodium silicate (Na2SiO3) to NaOH was 2.33, and the ratio of alkaline liquid to binder was 0.35. Based on experimental data collected from the literature, an evolutionary-based algorithm (GEP) was proposed to develop new predictive models for estimating the compressive strength of GPC containing bacteria. Data were classified into training and testing sets to obtain a closed-form solution using GEP. Independent variables for the model were the constituent materials of GPC, such as FA, MK, SF, and Bacillus bacteria. A total of six GEP formulations were developed for predicting the compressive strength of bacteria-incorporated GPC obtained at 1, 3, 7, 28, 56, and 90 days of curing. 80% and 20% of the data were used for training and testing the models, respectively. R2 values in the range of 0.9747 and 0.9950 (including train and test dataset) were obtained for the concrete samples, which showed that GEP can be used to predict the compressive strength of GPC containing bacteria with minimal error. Moreover, the GEP models were in good agreement with the experimental datasets and were robust and reliable. The models developed could serve as a tool for concrete constructors using geopolymers within the framework of this research.

• International journal of thyroidology
• /
• v.10 no.1
• /
• pp.14-23
• /
• 2017

Background and Objectives: To evaluate and compare the diagnostic performances of grayscale ultrasound (US) and quantitative parameters obtained from texture analysis of grayscale US and elastography images in evaluating patients with diffuse thyroid disease (DTD). Materials and Methods: From September to December 2012, 113 patients (mean age, $43.4{\pm}10.7years$) who had undergone preoperative staging US and elastography were included in this study. Assessment of the thyroid parenchyma for the diagnosis of DTD was made if US features suggestive of DTD were present. Nine histogram parameters were obtained from the grayscale US and elastography images, from which 'grayscale index' and 'elastography index' were calculated. Diagnostic performances of grayscale US, texture analysis using grayscale US and elastography were calculated and compared. Results: Of the 113 patients, 85 (75.2%) patients were negative for DTD and 28 (24.8%) were positive for DTD on pathology. The presence of US features suggestive of DTD showed significantly higher rates of DTD on pathology, 60.7% to 8.2% (p<0.001). Specificity, accuracy, and positive predictive value was highest in US features, 91.8%, 84.1%, and 87.6%, respectively (all ps<0.05). Grayscale index showed higher sensitivity and negative predictive value (NPV) than US features. All diagnostic performances were higher for grayscale index than the elastography index. Area under the curve of US features was the highest, 0.762, but without significant differences to grayscale index or mean of elastography (all ps>0.05). Conclusion: Diagnostic performances were the highest for grayscale US features in diagnosis of DTD. Grayscale index may be used as a complementary tool to US features for improving sensitivity and NPV.

• Journal of The Korean Association of Information Education
• /
• v.24 no.6
• /
• pp.597-606
• /
• 2020

Recently, as students' programming classes are being conducted online, interest in factors that can lead to the success of online programming classes is also increasing. Therefore, in this study, online programming classes were conducted for specialized high school students using a web-based simulation programming tool through TinkerCad. In these online programming classes, students' self-regulation ability and learning flow were set as variables that influence both learning engagement and teaching presence, and the predictive power of each was analyzed. As a result, it was found that both self-regulation ability and learning flow were predictive variables for learning engagement and teaching presence, and that learning flow played a mediating role between self-regulation ability, learning engagement, and teaching presence. This study is meaningful in that it suggested that self-regulation ability and learning flow should be considered more meaningfully in online programming classes, and a practical strategy for this is presented.

• Steel and Composite Structures
• /
• v.45 no.2
• /
• pp.205-218
• /
• 2022

Proper calculation of splitting tensile strength (STS) of concrete has been a crucial task, due to the wide use of concrete in the construction sector. Following many recent studies that have proposed various predictive models for this aim, this study suggests and tests the functionality of three hybrid models in predicting the STS from the characteristics of the mixture components including cement compressive strength, cement tensile strength, curing age, the maximum size of the crushed stone, stone powder content, sand fine modulus, water to binder ratio, and the ratio of sand. A multi-layer perceptron (MLP) neural network incorporates invasive weed optimization (IWO), cuttlefish optimization algorithm (CFOA), and electrostatic discharge algorithm (ESDA) which are among the newest optimization techniques. A dataset from the earlier literature is used for exploring and extrapolating the STS behavior. The results acquired from several accuracy criteria demonstrated a nice learning capability for all three hybrid models viz. IWO-MLP, CFOA-MLP, and ESDA-MLP. Also in the prediction phase, the prediction products were in a promising agreement (above 88%) with experimental results. However, a comparative look revealed the ESDA-MLP as the most accurate predictor. Considering mean absolute percentage error (MAPE) index, the error of ESDA-MLP was 9.05%, while the corresponding value for IWO-MLP and CFOA-MLP was 9.17 and 13.97%, respectively. Since the combination of MLP and ESDA can be an effective tool for optimizing the concrete mixture toward a desirable STS, the last part of this study is dedicated to extracting a predictive formula from this model.

• Journal of Trauma and Injury
• /
• v.37 no.2
• /
• pp.124-131
• /
• 2024

Purpose: The aim of this study was to utilize the American College of Surgeons Trauma Quality Improvement Program (TQIP) database to identify risk factors associated with developing acute compartment syndrome (ACS) following lower extremity fractures. Specifically, a nomogram of variables was constructed in order to propose a risk calculator for ACS following lower extremity trauma. Methods: A large retrospective case-control study was conducted using the TQIP database to identify risk factors associated with developing ACS following lower extremity fractures. Multivariable regression was used to identify significant risk factors and subsequently, these variables were implemented in a nomogram to develop a predictive model for developing ACS. Results: Novel risk factors identified include venous thromboembolism prophylaxis type particularly unfractionated heparin (odds ratio [OR], 2.67; 95% confidence interval [CI], 2.33-3.05; P<0.001), blood product transfusions (blood per unit: OR 1.13 [95% CI, 1.09-1.18], P<0.001; platelets per unit: OR 1.16 [95% CI, 1.09-1.24], P<0.001; cryoprecipitate per unit: OR 1.13 [95% CI, 1.04-1.22], P=0.003). Conclusions: This study provides evidence to believe that heparin use and blood product transfusions may be additional risk factors to evaluate when considering methods of risk stratification of lower extremity ACS. We propose a risk calculator using previously elucidated risk factors, as well as the risk factors demonstrated in this study. Our nomogram-based risk calculator is a tool that will aid in screening for high-risk patients for ACS and help in clinical decision-making.

• Clinical Endoscopy
• /
• v.56 no.3
• /
• pp.325-332
• /
• 2023

Background/Aims: Administrative databases provide valuable information for large-cohort studies. This study aimed to evaluate the diagnostic accuracy of an administrative database for resected gastric adenomas. Methods: Data of patients who underwent endoscopic resection for benign gastric lesions were collected from three hospitals. Gastric adenoma cases were identified in the hospital database using International Classification of Diseases (ICD) 10-codes. The non-adenoma group included patients without gastric adenoma codes. The diagnostic accuracy for gastric adenoma was analyzed based on the pathological reports of the resected specimen. Results: Among 5,095 endoscopic resections with codes for benign gastric lesions, 3,909 patients were included in the analysis. Among them, 2,831 and 1,078 patients were allocated to the adenoma and non-adenoma groups, respectively. Regarding the overall diagnosis of gastric adenoma with ICD-10 codes, the sensitivity, specificity, positive predictive value, and negative predictive value were 98.7%, 88.5%, 95.2%, and 96.8%, respectively. There were no significant differences in these parameters between the tertiary and secondary centers. Conclusions: Administrative codes of gastric adenoma, according to ICD-10 codes, showed good accuracy and can serve as a useful tool to study prognosis of these patients in real-world data studies in the future.

• Korean Journal of Radiology
• /
• v.20 no.1
• /
• pp.158-165
• /
• 2019

Objective: Recent studies demonstrated that core needle biopsy (CNB) can effectively reduce the possibility of inconclusive results and prevent unnecessary diagnostic surgery. However, the effectiveness of CNB in patients with suspicious thyroid nodules has not been fully evaluated. This prospective study aimed to determine the potential of CNB to assess thyroid nodules with suspicious ultrasound (US) features. Materials and Methods: Patients undergoing CNB for thyroid nodules with suspicious features on US were enrolled between May and August 2016. Diagnostic performance and the incidence of non-diagnostic results, inconclusive results, conclusive results, malignancy, unnecessary surgery, and complications were analyzed. Subgroup analysis according to nodule size was performed. The risk factors associated with inconclusive results were evaluated using multivariate logistic regression analysis. Results: A total of 93 patients (102 thyroid nodules) were evaluated. All samples obtained from CNB were adequate for diagnosis. Inconclusive results were seen in 12.7% of cases. The diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value for diagnosis of malignancy were 93.8%, 100%, 100%, 78.9%, and 95%, respectively. None of the patients underwent unnecessary surgery. The diagnostic performance was not significantly different according to nodule size. On multivariate logistic regression analysis, larger nodule size and shorter needle length were independent risk factors associated with inconclusive results. Conclusion: Samples obtained by CNB were sufficient for diagnosis in all cases and resulted in high diagnostic values and conclusive results in the evaluation of suspicious thyroid nodules. These findings indicated that CNB is a promising diagnostic tool for suspicious thyroid nodules.

• The Korean Journal of Nuclear Medicine
• /
• v.34 no.3
• /
• pp.234-242
• /
• 2000

Purpose: Tc-99m MIBI scintimammography has been validated as an useful non-invasive diagnostic tool for the primary breast cancer. But most studies have included small population of patients. We have experienced a large study population and investigated the diagnostic usefulness of Tc-99m MIBI scintimammography in detection of primary breast cancer and axillary lymph node metastasis. Materials and Methods: This study included 305 patients who underwent scintimammogtaphy for palpable breast masses or abnormal radiologic findings. Tc-99m MIBI scintimammography was performed 10 minutes after intravenous injection of 925 MBq of Tc-99m MIBI. If the early image revealed abnormal finding, 3 hour delayed image was also acquired. We calculated early and delayed lesion to non-lesion ratios (L/N). The pathologic diagnosis was obtained from surgical operation or FNAB and compared with the results of Tc-99m MIBI scintimammography. Results: Malignant breast diseases were 155 and benign ones were 150. Tc-99m MIBI scintimammography revealed 132 true positive, 23 false negative, 10 false positive, and 140 true negative cases. The sensitivity, specificity, positive predictive value and negative predictive value for the primary breast cancer detection were 85.2%, 93.4%, 92.9%, and 85.9%, respectively. The sensitivity, specificity, positive predictive and negative predictive values of Tc-99m MIBI scintimammography in detecting metastatic axillary lymph node involvement were 22%, 90.4%, 61.9% and 62.3%, respectively. Early L/N of malignant breast disease was significantly higher than that of benign one ($2.44{\pm}0.97\;vs\;1.94{\pm}0.78$, p=0.01). Delayed L/N had no significant difference between malignant and benign breast diseases ($1.94{\pm}0.52\;vs\;1.91{\pm}0.73$, p=0.43). Conclusion: Our study revealed that Tc-99m MIBI scintimammography was an useful diagnostic tool for the diagnosis of breast cancer. And early L/N ratio might provide complementary role in the detection of breast cancer. But the Tc-99m MIBI scintimammography had limited value in the detection of small breast cancer (less than 1 cm) and axillary lymph node metastasis.