A5038180509- Color perception and design
- Digital Marketing and Social Media
- Technology Adoption and User Behaviour
- Sentiment Analysis and Opinion Mining
- Human-Automation Interaction and Safety
- Vehicle Noise and Vibration Control
- Ergonomics and Musculoskeletal Disorders
- Multisensory perception and integration
- Consumer Behavior in Brand Consumption and Identification
- Consumer Market Behavior and Pricing
- Noise Effects and Management
- Anomaly Detection Techniques and Applications
- Safety Warnings and Signage
- Advanced Sensor and Energy Harvesting Materials
- Consumer Retail Behavior Studies
- Color Science and Applications
- Tactile and Sensory Interactions
- Forensic Anthropology and Bioarchaeology Studies
- Gas Sensing Nanomaterials and Sensors
- Sensory Analysis and Statistical Methods
- Effects of Vibration on Health
- Time Series Analysis and Forecasting
- Technology and Data Analysis
- Advanced Text Analysis Techniques
- Dermatoglyphics and Human Traits
Dongduk Women's University
2021-2025
Sungkyunkwan University
2007-2023
Korea Advanced Institute of Science and Technology
2011-2022
Seoul National University
1998-2021
Sungkyul University
2019-2020
North Carolina State University
2019
Korea Institute of Science and Technology
2018
Kyung Hee University
2012
Kangbuk Samsung Hospital
2012
Deep learning models are efficient in the features that assist understanding complex patterns precisely. This study proposed a computerized process of classifying skin disease through deep based MobileNet V2 and Long Short Term Memory (LSTM). The model proved to be with better accuracy can work on lightweight computational devices. is maintaining stateful information for precise predictions. A grey-level co-occurrence matrix used assessing progress diseased growth. performance has been...
Biomedical images contain a huge number of sensor measurements that can provide disease characteristics. Computer-assisted analysis such parameters aids in the early detection disease, and as result medical professionals quickly selecting appropriate medications. Human Activity Recognition, abbreviated 'HAR', is prediction common human measurements, which consist movements walking, running, drinking, cooking, etc. It extremely advantageous for services sphere care, fitness trackers, senior...
The present study aims to compare and analyze the performance of two tokenizers, Mecab-Ko SentencePiece, in context natural language processing for sentiment analysis. adopts a comparative approach, employing five algorithms - Naive Bayes (NB), k-Nearest Neighbor (kNN), Support Vector Machine (SVM), Artificial Neural Networks (ANN), Long Short-Term Memory Recurrent (LSTM-RNN) evaluate each tokenizer. was assessed based on four widely used metrics field, accuracy, precision, recall, F1-score....
Sitting on a chair in an awkward posture or sitting for long period of time is risk factor musculoskeletal disorders. A postural habit that has been formed cannot be changed easily. It important to form proper from childhood as the lumbar disease during caused by their improper most likely recur. Thus, there need monitoring system classifies children’s postures. The purpose this paper develop classifying postures children using machine learning algorithms. convolutional neural network (CNN)...
Recently, its becomes easy to track down the data due availability in a large number. Although for management, processing, and obtainability, cloud computing is considered well-known approach organizational development on internet. Despite many advantages, has still numerous security challenges that can affect big-data usage computing. To find issues/challenges are faced by software vendors' organizations we conducted systematic literature review (SLR) through which have out 103 relevant...
The water solubility of molecules is one the most important properties in various chemical and medical research fields. Recently, machine learning-based methods for predicting molecular properties, including solubility, have been extensively studied due to advantage effectively reducing computational costs. Although made significant advances predictive performance, existing were still lacking interpreting predicted results. Therefore, we propose a novel multi-order graph attention network...
Breast cancer encompasses various subtypes with distinct prognoses, necessitating accurate stratification methods. Current techniques rely on quantifying gene expression in limited subsets. Given the complexity of breast tissues, effective detection and classification is crucial medical imaging. This study introduces a novel method, MPa-DCAE, which uses multi-patch-based deep convolutional auto-encoder (DCAE) framework combined VGG19 to detect classify histopathology images. The proposed...
The most used self-assessment method for assessing driving style is the Multidimensional Driving Style Inventory (MDSI). This study aims to adapt MDSI Korean drivers (MDSI-K) and confirm eight-factor structure of original version. Six hundred forty aged 20-70 agreed participate in this study. All participants had at least one year experience. Confirmatory factor analysis was performed determine whether fit structure, goodness-of-fit values were not acceptable. Through correlation analysis,...
Abstract Healable conductive materials have received considerable attention. However, their practical applications are impeded by low electrical conductivity and irreversible degradation after breaking/healing cycles. Here we report a highly completely reversible electron tunneling-assisted percolation network of silver nanosatellite particles for putty-like moldable healable nanocomposites. The densely uniformly distributed with bimodal size distribution generated the radical reactive...
The dependence of the electrical resistance on materials' geometry determines performance conductive nanocomposites. Here, we report invariable a nanocomposite over 30% strain. This is enabled by in situ-generated hierarchically structured silver nanosatellite particles, realizing short interparticle distance (4.37 nm) stretchable silicone rubber matrix. Furthermore, barrier height tuned to be negligible matching electron affinity work function silver. stretching results flow without...
Ensembling is a powerful technique to obtain the most accurate results. In some cases, large number of learners in ensemble learning mostly increases both computational load during test phase and error rate. To solve this problem, paper we propose an Ensemble Reduced Deep Regression (ERDeR) model, which combination Regressions (DRs), shrinkage methods, approaches. The framework proposed model contains three phases. first includes base regressions parallel DRs are used as learners. role these...