• 전자공학회논문지CI
• /
• 제45권6호
• /
• pp.179-186
• /
• 2008

본 논문에서는 베이지안 알고리즘이 불균형 데이터의 학습 시 나타나는 현상을 분석하고 성능 평가 방법을 비교하였다. 사전 데이터 분포를 가정하고 불균형 데이터 비율과 분류 복잡도에 따라 발생된 분류 문제에 대해 베이지안 학습을 수행하였다. 실험 결과는 ROC(Receiver Operator Characteristic)와 PR(Precision-Recall) 평가 방법의 AUC(Area Under the Curve)를 계사하여 불균형 데이터 비율과 분류 복잡도에 따라 분석되었다. 비교 분석에서 불균형 비율은 기 수행된 연구 결과와 같이 베이지안 학습에 영향을 주었으며, 높은 분류 복잡도로부터 나타나는 데이터 중복은 학습 성능을 방해하는 요인으로 확인되었다. PR 평가의 AUC는 높은 분류 복잡도와 높은 불균형 데이터 비율에서 ROC 평가의 AUC보다 학습 성능의 차이가 크게 나타났다. 그러나 낮은 분류 복잡도와 낮은 불균형 데이터 비율의 문제에서 두 측정 방법의 학습 성능의 차이는 미비하거나 비슷하였다. 이러한 결과로부터 PR 평가의 AUC는 클래스 불균형 문제의 학습 모델의 설계와 오분류 비용을 고려한 최적의 학습기를 결정하는데 도움을 줄 수 있다.

• 한국산학기술학회논문지
• /
• 제21권6호
• /
• pp.374-382
• /
• 2020

본 연구에서 저장탄약 신뢰성평가(ASRP: Ammunition Stockpile Reliability Program)의 데이터 특성을 고려하여 입력변수를 줄이는 정규화기법을 제안함으로써 분류성능의 저하 없이 저장탄약 신뢰성분류 인경신경망모델의 학습 속도향상을 목표로 하였다. 탄약의 성능에 대한 기준은 국방규격(KDS: Korea Defense Specification)과 저장탄약 시험절차서(ASTP: Ammunition Stockpile reliability Test Procedure)에 규정되어 있으며, 평가결과 데이터는 이산형과 연속형 데이터가 복합적으로 구성되어 있다. 이러한 저장탄약 신뢰성평가의 데이터 특성을 고려하여 입력변수는 로트 추정 불량률(estimated lot percent nonconforming) 또는 고장률로 정규화 하였다. 또한 입력변수의 unitary hypercube를 유지하기 위하여 최소-최대 정규화를 2차로 수행하는 2단계 정규화 기법을 제안하였다. 제안된 2단계 정규화 기법은 저장탄약 신뢰성평가 데이터를 이용하여 비교한 결과 최소-최대 정규화와 유사하게 AUC(Area Under the ROC Curve)는 0.95 이상이었으며 학습속도는 학습 데이터 수와 은닉 계층의 노드 수에 따라 1.74 ~ 1.99 배 향상되었다.

• 전기학회논문지
• /
• 제67권7호
• /
• pp.928-933
• /
• 2018

Nowadays, gastropathy is a common disease. As endoscopic equipment are developed and used widely, it is possible to provide a large number of endoscopy images. Computer-aided Diagnosis (CADx) systems aim at helping physicians to identify possibly malignant abnormalities more accurately. In this paper, we present a CADx system to detect and classify the abnormalities of gastric lesions which include bleeding, ulcer, neuroendocrine tumor and cancer. We used an Inception module based deep learning model. And we used data augmentation for learning. Our preliminary results demonstrated promising potential for automatically labeled region of interest for endoscopy doctors to focus on abnormal lesions for subsequent targeted biopsy, with Az values of Receiver Operating Characteristic(ROC) curve was 0.83. The proposed CADx system showed reliable performance.

• 대한임베디드공학회논문지
• /
• 제18권5호
• /
• pp.225-231
• /
• 2023

With the recent surge in the autonomous driving market, the significance of lane detection technology has escalated. Lane detection plays a pivotal role in autonomous driving systems by identifying lanes to ensure safe vehicle operation. Traditional lane detection models rely on engineers manually extracting lane features from predefined environments. However, real-world road conditions present diverse challenges, hampering the engineers' ability to extract adaptable lane features, resulting in limited performance. Consequently, recent research has focused on developing deep learning based lane detection models to extract lane features directly from data. In this paper, we classify lane detection models into four categories: cluster-based, curve-based, information propagation-based, and anchor-based methods. We conduct an extensive analysis of the strengths and weaknesses of each approach, evaluate the model's performance on an embedded board, and assess their practicality and effectiveness. Based on our findings, we propose future research directions and potential enhancements.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권9호
• /
• pp.2334-2347
• /
• 2023

Metaverse services generate text data, data of ubiquitous computing, in real-time to analyze user emotions. Analysis of user emotions is an important task in metaverse services. This study aims to classify user sentiments using deep learning and pre-trained language models based on the transformer structure. Previous studies collected data from a single platform, whereas the current study incorporated the review data as "Metaverse" keyword from the YouTube and Google Play Store platforms for general utilization. As a result, the Bidirectional Encoder Representations from Transformers (BERT) and Robustly optimized BERT approach (RoBERTa) models using the soft voting mechanism achieved a highest accuracy of 88.57%. In addition, the area under the curve (AUC) score of the ensemble model comprising RoBERTa, BERT, and A Lite BERT (ALBERT) was 0.9458. The results demonstrate that the ensemble combined with the RoBERTa model exhibits good performance. Therefore, the RoBERTa model can be applied on platforms that provide metaverse services. The findings contribute to the advancement of natural language processing techniques in metaverse services, which are increasingly important in digital platforms and virtual environments. Overall, this study provides empirical evidence that sentiment analysis using deep learning and pre-trained language models is a promising approach to improving user experiences in metaverse services.

• Imaging Science in Dentistry
• /
• 제54권1호
• /
• pp.33-41
• /
• 2024

Purpose: The aims of this study were to create a deep learning model to distinguish between nasopalatine duct cysts (NDCs), radicular cysts, and no-lesions (normal) in the midline region of the anterior maxilla on panoramic radiographs and to compare its performance with that of dental residents. Materials and Methods: One hundred patients with a confirmed diagnosis of NDC (53 men, 47 women; average age, 44.6±16.5 years), 100 with radicular cysts (49 men, 51 women; average age, 47.5±16.4 years), and 100 with normal groups (56 men, 44 women; average age, 34.4±14.6 years) were enrolled in this study. Cases were randomly assigned to the training datasets (80%) and the test dataset (20%). Then, 20% of the training data were randomly assigned as validation data. A learning model was created using a customized DetectNet built in Digits version 5.0 (NVIDIA, Santa Clara, USA). The performance of the deep learning system was assessed and compared with that of two dental residents. Results: The performance of the deep learning system was superior to that of the dental residents except for the recall of radicular cysts. The areas under the curve (AUCs) for NDCs and radicular cysts in the deep learning system were significantly higher than those of the dental residents. The results for the dental residents revealed a significant difference in AUC between NDCs and normal groups. Conclusion: This study showed superior performance in detecting NDCs and radicular cysts and in distinguishing between these lesions and normal groups.

• 한국컴퓨터정보학회논문지
• /
• 제26권7호
• /
• pp.9-17
• /
• 2021

본 연구는 대장암 환자의 치료방법 중 하나인 항암화학요법을 분류할 수 있는 시스템인 CDSS연구의 일환으로 시행되었다. 대장암 치료에서 환자의 상태에 맞는 항암화학요법의 선택은 환자의 생존 기간과 직결되기 때문에 매우 중요하다. 따라서 본 연구에서는 대장암 환자의 개인적, 병리학적 특성을 사용해 기저 모델, 병리학적 모델, 그리고 환자의 두 가지 특성을 모두 사용한 결합 모델을 만들어 머신러닝 알고리즘으로 항암화학요법을 분류하였다. Top-n Accuracy와 ROC 곡선, AUC로 모델의 예측 정확도를 비교한 결과, 결합 모델에서 가장 우수한 예측 정확도를 보였으며, LGBM 알고리즘의 성능이 가장 우수한 것을 알 수 있었다. 본 연구에서는 머신러닝 알고리즘을 이용해 환자 특성별 모델을 분류함으로써 환자의 상태에 맞는 항암화학요법 분류 모델을 구축하였다. 향후 연구에서 본 연구 결과를 기초한다면 더 좋은 성능의 항암화학요법 분류 모델을 만들어 CDSS 연구에 도움이 될 것이다.

• Tuberculosis and Respiratory Diseases
• /
• 제86권3호
• /
• pp.226-233
• /
• 2023

Background: Inactive or old, healed tuberculosis (TB) on chest radiograph (CR) is often found in high TB incidence countries, and to avoid unnecessary evaluation and medication, differentiation from active TB is important. This study develops a deep learning (DL) model to estimate activity in a single chest radiographic analysis. Methods: A total of 3,824 active TB CRs from 511 individuals and 2,277 inactive TB CRs from 558 individuals were retrospectively collected. A pretrained convolutional neural network was fine-tuned to classify active and inactive TB. The model was pretrained with 8,964 pneumonia and 8,525 normal cases from the National Institute of Health (NIH) dataset. During the pretraining phase, the DL model learns the following tasks: pneumonia vs. normal, pneumonia vs. active TB, and active TB vs. normal. The performance of the DL model was validated using three external datasets. Receiver operating characteristic analyses were performed to evaluate the diagnostic performance to determine active TB by DL model and radiologists. Sensitivities and specificities for determining active TB were evaluated for both the DL model and radiologists. Results: The performance of the DL model showed area under the curve (AUC) values of 0.980 in internal validation, and 0.815 and 0.887 in external validation. The AUC values for the DL model, thoracic radiologist, and general radiologist, evaluated using one of the external validation datasets, were 0.815, 0.871, and 0.811, respectively. Conclusion: This DL-based algorithm showed potential as an effective diagnostic tool to identify TB activity, and could be useful for the follow-up of patients with inactive TB in high TB burden countries.

• Journal of Korean Neurosurgical Society
• /
• 제62권4호
• /
• pp.442-449
• /
• 2019

Objective : Bone mineral density (BMD) is an important consideration during fusion surgery. Although dual X-ray absorptiometry is considered as the gold standard for assessing BMD, quantitative computed tomography (QCT) provides more accurate data in spine osteoporosis. However, QCT has the disadvantage of additional radiation hazard and cost. The present study was to demonstrate the utility of artificial intelligence and machine learning algorithm for assessing osteoporosis using Hounsfield units (HU) of preoperative lumbar CT coupling with data of QCT. Methods : We reviewed 70 patients undergoing both QCT and conventional lumbar CT for spine surgery. The T-scores of 198 lumbar vertebra was assessed in QCT and the HU of vertebral body at the same level were measured in conventional CT by the picture archiving and communication system (PACS) system. A multiple regression algorithm was applied to predict the T-score using three independent variables (age, sex, and HU of vertebral body on conventional CT) coupling with T-score of QCT. Next, a logistic regression algorithm was applied to predict osteoporotic or non-osteoporotic vertebra. The Tensor flow and Python were used as the machine learning tools. The Tensor flow user interface developed in our institute was used for easy code generation. Results : The predictive model with multiple regression algorithm estimated similar T-scores with data of QCT. HU demonstrates the similar results as QCT without the discordance in only one non-osteoporotic vertebra that indicated osteoporosis. From the training set, the predictive model classified the lumbar vertebra into two groups (osteoporotic vs. non-osteoporotic spine) with 88.0% accuracy. In a test set of 40 vertebrae, classification accuracy was 92.5% when the learning rate was 0.0001 (precision, 0.939; recall, 0.969; F1 score, 0.954; area under the curve, 0.900). Conclusion : This study is a simple machine learning model applicable in the spine research field. The machine learning model can predict the T-score and osteoporotic vertebrae solely by measuring the HU of conventional CT, and this would help spine surgeons not to under-estimate the osteoporotic spine preoperatively. If applied to a bigger data set, we believe the predictive accuracy of our model will further increase. We propose that machine learning is an important modality of the medical research field.

• Geomechanics and Engineering
• /
• 제36권5호
• /
• 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.