Article
Computer Science, Information Systems
Zichen Zhang, Shifei Ding, Yuting Sun
Summary: This paper introduces a new method called multiple birth support vector regression (MBSVR), which constructs the regressor from multiple hyperplanes obtained by solving small quadratic programming problems, aiming for faster computation and better fitting precision.
INFORMATION SCIENCES
(2021)
Article
Environmental Sciences
Zhihao Wang, Alexander Brenning
Summary: Using active learning with uncertainty sampling can reduce the time and cost needed by experts under limited data conditions, improve model performance, and is particularly suitable for emergency response settings and landslide susceptibility modeling.
Article
Computer Science, Artificial Intelligence
Liming Liu, Maoxiang Chu, Rongfen Gong, Li Zhang
Summary: The improved nonparallel support vector machine (INPSVM) proposed in this article inherits the advantages of nonparallel support vector machine (NPSVM) while also offering incomparable benefits over twin support vector machine (TSVM). INPSVM effectively eliminates noise effects and achieves higher classification accuracy for both linear and nonlinear datasets compared to other algorithms. Experimental results demonstrate the superior efficiency, accuracy, and robustness of INPSVM.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Chen Ding, Tian-Yi Bao, He-Liang Huang
Summary: The study proposes a quantum-inspired classical algorithm for LS-SVM, utilizing an improved sampling technique for classification. The theoretical analysis indicates that the algorithm can achieve classification with logarithmic runtime for low-rank, low-condition number, and high-dimensional data matrices.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Information Systems
Ceren Atik, Recep Alp Kut, Reyat Yilmaz, Derya Birant
Summary: This paper proposes a novel method called support vector machine chains (SVMC) that involves chaining together multiple SVM classifiers in a special structure, decrementing one feature at each stage. The paper also introduces a new voting mechanism called tournament voting, where classifiers' outputs compete in groups and the winning class label of the final round is assigned as the prediction. Experimental results show that SVMC outperforms SVM in terms of accuracy and achieves a 6.88% improvement over state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Matteo Avolio, Antonio Fuduli
Summary: This paper introduces a novel approach for binary multiple instance learning classification, combining the strengths of SVM and PSVM, aiming to discriminate between positive and negative instances by generating a hyperplane placed in the middle between two parallel hyperplanes.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Information Systems
Lili Zhu, Petros Spachos
Summary: Food quality and safety are crucial for human health and social stability. This study proposed a mobile visual system to grade bananas, achieving high accuracy rates in the grading process. The complex process of ensuring food quality involves all stages from cultivation to consumption.
INTERNET OF THINGS
(2021)
Review
Agriculture, Multidisciplinary
Zhi Hong Kok, Abdul Rashid Mohamed Shariff, Meftah Salem M. Alfatni, Siti Khairunniza-Bejo
Summary: The Support Vector Machine (SVM) shows excellent performance in precision agriculture (PA), with comparisons to other machine learning algorithms highlighting its strengths and weaknesses in model performance and characteristics.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Computer Science, Artificial Intelligence
Zongmin Liu, Yitian Xu
Summary: In this paper, a novel multi-task nonparallel support vector machine (MTNPSVM) is proposed, which effectively avoids matrix inversion operation and takes full advantage of the kernel trick by introducing epsilon-insensitive loss instead of square loss. The alternating direction method of multipliers (ADMM) is employed to improve computational efficiency, and the properties and sensitivity of the model parameters are further explored.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Chunling Lou, Xijiong Xie
Summary: Two novel multi-view intuitionistic fuzzy support vector machines with insensitive pinball loss are proposed in this paper, which can handle general multi-view classification problems and be robust to noisy data. The pinball loss is incorporated into the multi-view learning to maximize the quantization distance. Intuitionistic fuzzy score is introduced to assign weights to the multi-view samples to effectively utilize multi-view information.
Article
Environmental Sciences
Kouao Laurent Kouadio, Loukou Nicolas Kouame, Coulibaly Drissa, Binbin Mi, Kouamelan Serge Kouamelan, Serge Pacome Deguine Gnoleba, Hongyu Zhang, Jianghai Xia
Summary: This study applied support vector machines (SVMs) to predict flow rates in groundwater exploration, aiming to minimize unsuccessful drillings. The SVM models achieved prediction accuracies of 77% and 83% on multiclass and binary datasets, respectively. The use of optimal polynomial and radial basis function kernels resulted in higher accuracies of 81.61% and 87.36%. Learning curves showed that larger training data could improve prediction performance on the multiclass dataset, but not necessarily on the binary dataset.
WATER RESOURCES RESEARCH
(2022)
Article
Critical Care Medicine
Yuzhuo Zhao, Lijing Jia, Ruiqi Jia, Hui Han, Cong Feng, Xueyan Li, Zijian Wei, Hongxin Wang, Heng Zhang, Shuxiao Pan, Jiaming Wang, Xin Guo, Zheyuan Yu, Xiucheng Li, Zhaohong Wang, Wei Chen, Jing Li, Tanshi Li
Summary: Early warning prediction of traumatic hemorrhagic shock can reduce patient mortality and morbidity. Different models with varied feature sets were developed and validated using machine learning algorithms. Features in vital signs, routine blood, and blood gas analysis were found to be the most relevant to traumatic hemorrhagic shock. The model performed best when predicting within a 1-hour time window.
Article
Mathematics
Guvenc Arslan, Ugur Madran, Duygu Soyoglu
Summary: In this paper, a novel classification approach is proposed by introducing a new clustering method as an intermediate step to discover the structure of a data set and reduce its size. Experimental results show that the proposed method performs comparably to standard support vector machines.
Article
Biotechnology & Applied Microbiology
Ke-Fan Wang, Jing An, Zhen Wei, Can Cui, Xiang-Hua Ma, Chao Ma, Han-Qiu Bao
Summary: In this paper, a novel imbalance classification method based on deep learning and fuzzy support vector machine, named DFSVM, is proposed. The method utilizes a deep neural network to obtain an embedding representation of the data and performs oversampling in the embedding space to address the data imbalance issue. Furthermore, a fuzzy support vector machine is used as the final classifier to improve the classification quality of minority classes.
FRONTIERS IN BIOENGINEERING AND BIOTECHNOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Ganesan Kalaiarasi, Sureshbabu Maheswari
Summary: In this study, an effective classification of hyperspectral images was modeled and simulated with the proximal support vector machine (PSVM) by integrating them with the deep learning approach. The new deep PSVM classifiers, designed to handle the complexity, discrepancies, and irregularities in traditional hyperspectral image classifiers, showed better classification accuracy compared to other techniques.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Ophthalmology
Tyler Hyungtaek Rim, Aaron Y. Lee, Daniel S. Ting, Kelvin Teo, Bjorn Kaijun Betzler, Zhen Ling Teo, Tea Keun Yoo, Geunyoung Lee, Youngnam Kim, Andrew C. Lin, Seong Eun Kim, Yih Chung Tham, Sung Soo Kim, Ching-Yu Cheng, Tien Yin Wong, Chui Ming Gemmy Cheung
Summary: The study evaluated the generalization ability of a DL model trained on Korean data to an American data set, and confirmed its accuracy through expert grading.
BRITISH JOURNAL OF OPHTHALMOLOGY
(2021)
Article
Ophthalmology
Tyler Hyungtaek Rim, Aaron Yuntai Lee, Daniel S. Ting, Kelvin Yi Chong Teo, Hee Seung Yang, Hyeonmin Kim, Geunyoung Lee, Zhen Ling Teo, Alvin Teo Wei Jun, Kengo Takahashi, Tea Keun Yoo, Sung Eun Kim, Yasuo Yanagi, Ching-Yu Cheng, Sung Soo Kim, Tien Yin Wong, Chui Ming Gemmy Cheung
Summary: This study developed a computer-aided detection (CADe) algorithm using deep learning to automatically segment and classify ORL abnormalities in OCT images. The algorithm achieved high accuracy in both internal and external test sets, and successfully differentiated between normal and abnormal ORL.
BRITISH JOURNAL OF OPHTHALMOLOGY
(2022)
Article
Ophthalmology
Tae Keun Yoo, Ik Hee Ryu, Jin Kuk Kim, In Sik Lee
Summary: This study evaluated a deep learning model for estimating uncorrected refractive error using OCT images, showing potential for estimating refractive error and detecting high myopia. The model performed well in predicting spherical equivalent and detecting high myopia, highlighting inner retinal layers and steepened curvatures as characteristic features.
Article
Ophthalmology
Tae Keun Yoo, Bo Yi Kim, Hyun Kyo Jeong, Hong Kyu Kim, Donghyun Yang, Ik Hee Ryu
Summary: In this study, a deep learning approach was used to segment and evaluate central serous chorioretinopathy (CSC) using fundus photographs. The results showed that the proposed U-Net model based on pix2pix algorithm had good segmentation performance and could accurately detect SRF lesions. This method can be easily implemented in a web-based environment without prior coding skills or personal computing resources.
TRANSLATIONAL VISION SCIENCE & TECHNOLOGY
(2022)
Review
Ophthalmology
Aram You, Jin Kuk Kim, Ik Hee Ryu, Tae Keun Yoo
Summary: GAN has a wide range of applications in ophthalmology image domains, including segmentation, data augmentation, denoising, domain transfer, super-resolution, post-intervention prediction, and feature extraction. However, GAN also has limitations such as mode collapse, spatial deformities, unintended changes, and the generation of high-frequency noises and artifacts of checkerboard patterns.
Article
Computer Science, Interdisciplinary Applications
Tae Keun Yoo, Ik Hee Ryu, Jin Kuk Kim, In Sik Lee, Hong Kyu Kim
Summary: In this study, a deep learning approach was used to detect shallow anterior chamber depth (ACD) in angle-closure glaucoma (ACG) using fundus photographs. The characteristic features of shallow ACD, previously undetectable by conventional techniques, were visualized using CycleGAN-based feature maps. This research demonstrates the feasibility of using deep learning models to detect shallow ACD and improve screening for ACG.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2022)
Review
Ophthalmology
Tae Keun Yoo, Seung Min Lee, Hansang Lee, Eun Young Choi, Min Kim
Summary: Retropupillary iris fixation of Artisan Myopia IOLs appears to be a safe and effective surgical treatment option for correcting aphakia and IOL dislocation in patients with extremely high myopia, with no significant postoperative complications.
OPHTHALMOLOGY AND THERAPY
(2022)
Article
Ophthalmology
Juntae Kim, Ik Hee Ryu, Jin Kuk Kim, In Sik Lee, Hong Kyu Kim, Eoksoo Han, Tae Keun Yoo
Summary: This study developed machine learning models to identify high-risk patients for refractive regression based on preoperative data and fundus photography. The models showed good predictive performance, providing an efficient strategy for surgeons to assess the risk of myopic regression and make informed decisions.
GRAEFES ARCHIVE FOR CLINICAL AND EXPERIMENTAL OPHTHALMOLOGY
(2022)
Article
Medicine, General & Internal
Bo Ram Kim, Tae Keun Yoo, Hong Kyu Kim, Ik Hee Ryu, Jin Kuk Kim, In Sik Lee, Jung Soo Kim, Dong-Hyeok Shin, Young-Sang Kim, Bom Taeck Kim
Summary: This study used eye examinations to predict sarcopenia and help identify individuals at high risk early on. Factors such as decreased levator function, cataracts, and age-related macular degeneration were associated with increased risk of sarcopenia in men, while blepharoptosis and cataracts were associated with increased risk in women. The XGBoost technique showed promising results with AUCs of 0.746 and 0.762 in men and women, respectively.
Article
Multidisciplinary Sciences
Tae Keun Yoo, Seo Hee Kim, Min Kim, Christopher Seungkyu Lee, Suk Ho Byeon, Sung Soo Kim, Jinyoung Yeo, Eun Young Choi
Summary: In this study, a two-stage deep learning model was developed to predict the 1-year outcome of photodynamic therapy (PDT) for chronic central serous chorioretinopathy (CSC) using initial multimodal clinical data. The model achieved high accuracy in predicting the treatability of CSC and outperformed a domain-specific ResNet50 model. The deep features from fundus photographs had the greatest impact on the model's performance.
SCIENTIFIC REPORTS
(2022)
Editorial Material
Oncology
Ein Oh, Yong Hyun Kim, Ik Hee Ryu, Tae Keun Yoo
ANNALS OF TRANSLATIONAL MEDICINE
(2022)
Article
Ophthalmology
Hannuy Choi, Taein Kim, Su Jeong Kim, Beom Gi Sa, Ik Hee Ryu, In Sik Lee, Jin Kuk Kim, Eoksoo Han, Hong Kyu Kim, Tae Keun Yoo
Summary: Linear regression and machine learning models were developed to predict postoperative ACAs for ICL surgery. Surgeons can use these models to select the optimal ICL size and reduce the risk of ACA-related complications.
TRANSLATIONAL VISION SCIENCE & TECHNOLOGY
(2023)
Review
Ophthalmology
Taein Kim, Su Jeong Kim, Bo Young Lee, Hye Jin Cho, Beom Gi Sa, Ik Hee Ryu, Jin Kuk Kim, In Sik Lee, Eoksoo Han, Hyungsu Kim, Tae Keun Yoo
Summary: We developed a prediction model for postoperative intraocular collamer lens (ICL) vault using Heidelberg anterior segment optical coherence tomography (AS-OCT) device. The model showed strong correlation and low mean squared error, indicating accurate prediction of ICL vault.
Article
Computer Science, Interdisciplinary Applications
Joon Yul Choi, Hyungsu Kim, Jin Kuk Kim, In Sik Lee, Ik Hee Ryu, Jung Soo Kim, Tae Keun Yoo
Summary: This study aims to develop a deep learning model based on fundus photography (FP) to predict corneal curvature by categorizing corneas into different groups. The developed algorithm shows that FP can potentially be used in telemedicine to detect abnormal corneal curvatures.
MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING
(2023)
Article
Ophthalmology
Rishabh Jain, Tae Keun Yoo, Ik Hee Ryu, Joanna Song, Nitin Kolte, Ashiyana Nariani
Summary: In this study, a deep transfer learning model with adaptation training was used to predict refractive errors and corneal curvature using OCT images. The adaptation training improved the performance, and different types of OCT images showed the best performance in different tasks.
OPHTHALMOLOGY AND THERAPY
(2023)