• Journal of the Korean Society of Radiology
• /
• v.81 no.5
• /
• pp.1164-1174
• /
• 2020

Purpose To evaluate a deep learning model to predict recurrence of thyroid tumor using preoperative ultrasonography (US). Materials and Methods We included representative images from 229 US-based patients (male:female = 42:187; mean age, 49.6 years) who had been diagnosed with thyroid cancer on preoperative US and subsequently underwent thyroid surgery. After selecting each representative transverse or longitudinal US image, we created a data set from the resulting database of 898 images after augmentation. The Python 2.7.6 and Keras 2.1.5 framework for neural networks were used for deep learning with a convolutional neural network. We compared the clinical and histological features between patients with and without recurrence. The predictive performance of the deep learning model between groups was evaluated using receiver operating characteristic (ROC) analysis, and the area under the ROC curve served as a summary of the prognostic performance of the deep learning model to predict recurrent thyroid cancer. Results Tumor recurrence was noted in 49 (21.4%) among the 229 patients. Tumor size and multifocality varied significantly between the groups with and without recurrence (p < 0.05). The overall mean area under the curve (AUC) value of the deep learning model for prediction of recurrent thyroid cancer was 0.9 ± 0.06. The mean AUC value was 0.87 ± 0.03 in macrocarcinoma and 0.79 ± 0.16 in microcarcinoma. Conclusion A deep learning model for analysis of US images of thyroid cancer showed the possibility of predicting recurrence of thyroid cancer.

• The Journal of the Convergence on Culture Technology
• /
• v.6 no.4
• /
• pp.289-294
• /
• 2020

Peptic ulcer disease is a gastrointestinal disorder caused by Helicobacter pylori infection and the use of nonsteroid anti-inflammatory drugs. While many studies have been conducted to find the risk factors of peptic ulcers, there are no studies on the suggestion of peptic ulcer prediction models for Koreans. Therefore, the purpose of this study is to implement peptic ulcer prediction model using machine learning based on demographic information, obesity information, blood information, and nutritional information for middle-aged and elderly people. For model building, wrapper-based variable selection method and naive Bayes algorithm were used. The classification accuracy of the female prediction model was the area under the receiver operating characteristics curve (AUC) of 0.712, and males showed an AUC of 0.674, which is lower than that of females. These results can be used for prediction and prevention of peptic ulcers in the middle and elderly people.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.4
• /
• pp.337-346
• /
• 2021

Detection of abnormal signal generally can be done by using features of normal signals as main information because of data imbalance. This paper propose an efficient method for abnormal signal detection using parallel AutoEncoder (AE) which can use features of abnormal signals as well. The proposed Parallel AE (PAE) is composed of a normal and an abnormal reconstructors having identical AE structure and train features of normal and abnormal signals, respectively. The PAE can effectively solve the imbalanced data problem by sequentially training normal and abnormal data. For further detection performance improvement, additional binary classifier can be added to the PAE. Through experiments using public acoustic data, we obtain that the proposed PAE shows Area Under Curve (AUC) improvement of minimum 22 % at the expenses of training time increased by 1.31 ~ 1.61 times to the single AE. Furthermore, the PAE shows 93 % AUC improvement in detecting abnormal underwater acoustic signal when pre-trained PAE is transferred to train open underwater acoustic data.

• Journal of Veterinary Clinics
• /
• v.33 no.2
• /
• pp.97-101
• /
• 2016

In the field of clinical medicine, diagnostic accuracy studies refer to the degree of agreement between the index test and the reference standard for the discriminatory ability to identify a target disorder of interest in a patient. The receiver operating characteristic (ROC) curve offers a graphical display the trade-off between sensitivity and specificity at each cutoff for a diagnostic test and is useful in assigning the best cutoff for clinical use. In this end, the ROC curve analysis is a useful tool for estimating and comparing the accuracy of competing diagnostic tests. This paper reviews briefly the measures of diagnostic accuracy such as sensitivity, specificity, and area under the ROC curve (AUC) that is a summary measure for diagnostic accuracy across the spectrum of test results. In addition, the methods of creating an ROC curve in single diagnostic test with five-category discrete scale for disease classification from healthy individuals, meaningful interpretation of the AUC, and the applications of ROC methodology in clinical medicine to determine the optimal cutoff values have been discussed using a hypothetical example as an illustration.

• Journal of Korean Academy of Nursing
• /
• v.46 no.1
• /
• pp.59-68
• /
• 2016

Purpose: Increase in suicide rate for senior citizens which has become widespread in our society today. It is not a normal social phenomenon and is beyond the danger level. The contents of this study include Korean senior citizens' suicide related risk factors and warning signs, and the development of a simple Geriatric Suicide Risk Scale. Methods: This study is Methodological Research to verify reliability and validity of the Geriatric Suicide Risk Scale according to the tool development process suggested by Devellis (2012). Results: For predictive validity assessment, high suicide screening accuracy was showed with an Area under the ROC curve (AUC) of .93. For the optimal cutoff point of 11, sensitivity was 93.9%, and specificity, 75.7% which are excellence levels. Cross validity for assessment of generalization possibility showed the Area under the ROC curve (AUC) as .82 and in case of a cutoff point of 11, sensitivity was 73.7%, and specificity, 65.9%. Conclusion: When it comes to practical nursing, it is significant that the Korean Geriatric Suicide Risk Scale has high reliability and validity through adequate tool development and the tool assessment step to select degree of suicide risk of senior citizens. Also, it can be easily applied and does not take a long time to administer. Further, it can be used by health care personnel or the general public.

• Biomedical Science Letters
• /
• v.19 no.3
• /
• pp.233-238
• /
• 2013

Aspirin is still the mainstay of antiplatelet therapy in the cardiovascular and cerebrovascular disease. However, some patients are not responsive to the antithrombotic action of aspirin. The aim of this study was to assess the prevalence and clinical characteristics of aspirin resistance in patients with cerebral infarction. We tested platelet function in 557 patients who had been treated with aspirin in J general hospital. Platelet function was tested using the multiple electrode platelet aggregometry (MEA). Platelet reactivity was expressed as area under the aggregation curve (AUC, U) and >30 AUC was defined as aspirin resistance. Aspirin resistance was detected in 16.2% patients. There was not any significant differences in age, gender between aspirin resistance and aspirin sensitive patients. WBC was significantly higher in patients with aspirin resistance (P < .05). HDL-cholesterol was significantly higher in patients with aspirin sensitive (P < .05). Aspirin resistance was positive correlation with platelet count (r =.314, P =.003). The prevalence of aspirin resistance in cerebral infarction was 16.2%, and platelet count were related with aspirin resistance.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.3
• /
• pp.221-225
• /
• 2023

In this paper, we proposed an efficient method for rapid diagnosis of COVID-19 by voice. A novel Strided Convolution Separable Transformer (SC-SepTr) is proposed by modifying the conventional Separable Transformer (SepTr) for audio signal recognition. The proposed method reduces the memory and computational requirements to enable rapid diagnosis of COVID-19. As a result of experiments on Coswara, it was shown that the proposed method perform rapid diagnosis with guaranteeing Area Under the Curve (AUC) performance even for a relatively small amount of learning data.

• The Journal of the Convergence on Culture Technology
• /
• v.6 no.3
• /
• pp.457-462
• /
• 2020

Osteoporosis is a disease that occurs mainly in the elderly and increases the risk of fractures due to structural deterioration of bone mass and tissues. The purpose of this study are to assess the relationship between nutritional components and osteoporosis and to evaluate models for predicting osteoporosis based on nutrient components. In experimental method, association was performed using binary logistic regression, and predictive models were generated using the naive Bayes algorithm and variable subset selection methods. The analysis results for single variables indicated that food intake and vitamin B2 showed the highest value of the area under the receiver operating characteristic curve (AUC) for predicting osteoporosis in men. In women, monounsaturated fatty acids showed the highest AUC value. In prediction model of female osteoporosis, the models generated by the correlation based feature subset and wrapper based variable subset methods showed an AUC value of 0.662. In men, the model by the full variable obtained an AUC of 0.626, and in other male models, the predictive performance was very low in sensitivity and 1-specificity. The results of these studies are expected to be used as the basic information for the treatment and prevention of osteoporosis.

• Journal of Life Science
• /
• v.32 no.2
• /
• pp.79-85
• /
• 2022

Recently, many hangover cure products containing natural ingredients have been made available in the market that are effective for alcohol-related liver damage or for improved liver function. However, the cure for of liver damage or medication for improved liver function are different from hangover cure. Therefore, the efficacy hangover cure products needs to be verified. In this study, we investigated and compared the ameliorating effect of four commercially available hangover cure products on acute ethanol-induced hangover in Sprague - Dawley rats. The four samples were labeled as C, M, R, and S. The efficacy of the samples was evaluated based on the serum concentration and area under the curve (AUC) of blood ethanol and acetaldehyde concentrations to quantitatively assess the hangover cure effect. Ethanol administration to the rats significantly raised the serum alcohol and acetaldehyde levels. The Cmax reduction rates of ethanol for the samples C, M, R, and S were 5.9%, 3.1%, 8.4%, and 11.7%, and the AUC were 8.9%, 2.2%, 12.1%, and 19.6%, respectively, whereas the Cmax reduction rates of acetaldehyde were 14.2%, 15.2%, 28.2%, and 35.0%, and the AUC were 21.6%, 7.5%, 22.4%, and 29.9%, respectively. In conclusion, all samples showed a tendency to relieve hangover in the order of S, R, C, and M in terms of the ethanol concentration, but only sample S showed a statistically significant decrease in both Cmax and AUC for ethanol and acetaldehyde. These results suggest that an objective method for verifying the efficacy of hangover cure products is lacking.

• BMB Reports
• /
• v.46 no.1
• /
• pp.41-46
• /
• 2013

Identifying genes indispensable for an organism's life and their characteristics is one of the central questions in current biological research, and hence it would be helpful to develop computational approaches towards the prediction of essential genes. The performance of a predictor is usually measured by the area under the receiver operating characteristic curve (AUC). We propose a novel method by implementing genetic algorithms to maximize the partial AUC that is restricted to a specific interval of lower false positive rate (FPR), the region relevant to follow-up experimental validation. Our predictor uses various features based on sequence information, protein-protein interaction network topology, and gene expression profiles. A feature selection wrapper was developed to alleviate the over-fitting problem and to weigh each feature's relevance to prediction. We evaluated our method using the proteome of budding yeast. Our implementation of genetic algorithms maximizing the partial AUC below 0.05 or 0.10 of FPR outperformed other popular classification methods.