Article
Chemistry, Analytical
Imayanmosha Wahlang, Arnab Kumar Maji, Goutam Saha, Prasun Chakrabarti, Michal Jasinski, Zbigniew Leonowicz, Elzbieta Jasinska
Summary: Effective classification techniques using deep learning on brain MRI images with added attributes of gender and age show improved accuracy in diagnosing brain tumors. Age and gender play a key role in classification, and the proposed technique outperforms existing methods in most cases.
Article
Biology
Xueliang Zhu, Jie Ying, Haima Yang, Le Fu, Boyang Li, Bin Jiang
Summary: This study proposes a novel computerized method for accurately detecting deep myometrial invasion on MRI by utilizing the geometric feature LS and texture features in the ensemble model EPSVM. The results demonstrate that EPSVM outperforms commonly used classifiers in terms of accuracy, sensitivity, and specificity, and LS plays a significant role in DMI detection.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Psychiatry
Xin Guo, Wei Wang, Lijun Kang, Chang Shu, Hanpin Bai, Ning Tu, Lihong Bu, Yujun Gao, Gaohua Wang, Zhongchun Liu
Summary: This study investigates the relationship between adolescent depression and whole-brain network centrality abnormalities, and finds that changes in brain voxel-level functional connectivity may be a potential biomarker for adolescent depression.
FRONTIERS IN PSYCHIATRY
(2022)
Article
Computer Science, Artificial Intelligence
Shili Peng, Wenwu Wang, Yinli Chen, Xueling Zhong, Qinghua Hu
Summary: This article presents a new idea for addressing the challenge of unifying classification and regression in machine learning. It proposes converting the classification problem into a regression problem and using regression methods to solve key problems in classification. Experimental results demonstrate that the proposed method outperforms existing algorithms in terms of prediction accuracy and model uncertainty.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Qifan Yang, Huijuan Zhang, Jun Xia, Xiaoliang Zhang
Summary: Support vector machine (SVM) and convolutional neural network (CNN) are two widely used machine learning methods for brain low-grade glioma (LGG) magnetic resonance imaging (MRI) segmentation. While SVM models have shorter computation time but lower accuracy, the CNN model outperforms SVM in accuracy but with longer computation time.
QUANTITATIVE IMAGING IN MEDICINE AND SURGERY
(2021)
Article
Computer Science, Artificial Intelligence
Xiaochen Zhou, Xudong Wang
Summary: Fed-KSVM is a federated learning scheme designed for training low-memory-consumption kernel SVM models. By decomposing the training process into subproblems and using an incremental learning algorithm, it achieves reduced memory consumption on edge devices. Additionally, by constructing a global model after training the local models, the scheme reduces communication costs while maintaining high accuracy.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Information Systems
Tiep M. Hoang, Trung Q. Duong, Hoang Duong Tuan, Sangarapillai Lambotharan, Lajos Hanzo
Summary: This article presents a framework for converting wireless signals into structured datasets for detecting active eavesdropping attacks at the physical layer using machine learning algorithms.
Review
Operations Research & Management Science
M. Tanveer, T. Rajani, R. Rastogi, Y. H. Shao, M. A. Ganaie
Summary: TWSVM and TSVR are emerging machine learning techniques for classification and regression challenges. TWSVM classifies data points using two nonparallel hyperplanes, while TSVR is based on TWSVM and solves two SVM-type problems. Although there has been progress in research on these techniques, there is limited literature on the comparison of different variants of TSVR.
ANNALS OF OPERATIONS RESEARCH
(2022)
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
Management
Haimonti Dutta
Summary: In the era of big data, a scalable support vector machine (SVM) algorithm is an important tool for machine learning researchers. This paper presents a distributed algorithm, called the gossip-based subgradient (GADGET) SVM, for learning linear SVMs in the primal form. The algorithm can be executed locally on sites of a distributed system, and it has fast convergence speed and low message complexity. Empirical results show that the algorithm performs comparably to other state-of-the-art solvers.
MANAGEMENT SCIENCE
(2022)
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
Microbiology
Thomas J. Tewes, Mario Kerst, Frank Platte, Dirk P. Bockmuehl
Summary: In this study, predictive models based on Raman spectroscopy combined with support vector machines were developed for accurate identification of microorganisms. The method showed promising results for various applications such as medical diagnostics and the food industry.
Article
Computer Science, Information Systems
Massimo Donelli, Giuseppe Espa, Paola Feraco
Summary: Due to the complexity of brain structures, brain tumors have high mortality and disability rates, necessitating early diagnosis to minimize damage. While biopsies are the gold standard, they can be invasive and yield incorrect results due to intratumoral heterogeneity. Medical imaging procedures offer non-invasive and reproducible evaluation of the entire tumor. Radiomics, based on quantitative medical image analysis, aims to identify associations between tumor features, diagnosis, and patient prognosis to optimize treatments and improve survival rates. However, current radiomics techniques lack standardization in segmentation, feature extraction, and selection, requiring expert supervision for treatment decisions. This paper proposes a semi-automatic methodology utilizing binary texture recognition, growing area algorithms, and machine learning techniques to aid in the identification and segmentation of malignant tissues, with promising preliminary results.
Article
Radiology, Nuclear Medicine & Medical Imaging
Hong Peng, Jiaohua Huo, Bo Li, Yuanyuan Cui, Hao Zhang, Liang Zhang, Lin Ma
Summary: The study aimed to establish a model for classifying IDH status in gliomas based on multiparametric MRI, and found that radiomics features can identify IDH genotypes in gliomas.
JOURNAL OF MAGNETIC RESONANCE IMAGING
(2021)
Article
Automation & Control Systems
Gurparsad Singh Suri, Gurleen Kaur, Sara Moein
Summary: Schizophrenia is a mental disorder with unknown cause, and machine learning and artificial intelligence are used for its detection and monitoring by identifying the associations between symptoms and disease. Significant connections between brain regions and symptoms of schizophrenia have been found. Support vector machines, deep neural networks, and random forests can predict schizophrenia with high accuracy.
INTELLIGENT AUTOMATION AND SOFT COMPUTING
(2021)