Article
Computer Science, Interdisciplinary Applications
Santosh Kumar Sahu, Durga Prasad Mohapatra, Jitendra Kumar Rout, Kshira Sagar Sahoo, Ashish Kr Luhach
Summary: The study established a scalable framework for large-scale data processing and analytics, implemented, tuned, and evaluated popular classification methods using intrusion datasets, selected decision tree as the base classifier to study class imbalance issues.
Article
Computer Science, Artificial Intelligence
Ashok Kumar Bathla, Shally Bansal, Munish Kumar
Summary: The proposed method, OKC classifier, outperforms various well-established classification algorithms on imbalanced datasets without resampling.
Article
Ecology
Javad Seyedmohammadi, Ali Zeinadini, Mir Naser Navidi, Richard W. McDowell
Summary: To predict pistachio yield and manage soil and water resources sustainably, researchers developed models using classification and regression trees, k-nearest neighbors, support vector machines, and a new hybrid model. Through soil sampling and analysis from 124 pistachio orchards, certain soil properties were found to be important predictors of yield. The hybrid model combining support vector machines and the firefly meta-heuristic algorithm outperformed other models in capturing non-linear relationships. The results can help optimize soil management for increased pistachio yield and sustainable production.
ECOLOGICAL INFORMATICS
(2023)
Article
Remote Sensing
Bonggeun Song, Kyunghun Park
Summary: This study compared the accuracy of various machine learning techniques for the efficient detection and management of outdoor compost piles (OCPs) in South Korea. The findings showed high accuracy in predicting OCP location and area, especially with the random forest method. This research provides valuable data for improving OCP management strategies and water quality.
Article
Engineering, Electrical & Electronic
Yi-Feng Cheng, Dan Li, Shichang Chen, Gaofeng Wang
Summary: This study introduces a new approach for designing a wideband decoupling and matching network for two-element antenna arrays. Through the use of Even-Odd-Mode theory, the decoupling and matching problems are transformed into classical impedance matching problems, simplifying the network design process. By using a genetic algorithm to search for optimal circuit parameters, the proposed design demonstrates ultra-wide operating bandwidth, improved isolation, and high efficiency.
IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS
(2023)
Article
Environmental Sciences
Vikas Kumar Rana, Tallavajhala Maruthi Venkata Suryanarayana
Summary: This study explores the impact of urban landform clustering on land surface temperature, using different indices for analysis. The results show that the influence of different seasons and landscapes on land surface temperature varies. Additionally, the study evaluates the predictive performance of four machine learning models.
GEOCARTO INTERNATIONAL
(2022)
Article
Computer Science, Information Systems
Tran The Son, Chando Lee, Hoa Le -Minh, Nauman Aslam, Vuong Cong Dat
Summary: This paper proposes a simple and effective model for image-based malware classification using machine learning. The model does not use fixed-size square images or features extracted by image descriptor for training, but instead normalizes and reduces the dimension of the input images. The experiment shows that reducing the dimension of the images can decrease the complexity and training time of the model while maintaining classification accuracy.
JOURNAL OF INFORMATION SECURITY AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Fuad Ali Mohammed Al-Yarimi, Nabil Mohammed Ali Munassar, Mohammed Hasan Mohammed Bamashmos, Mohammed Yousef Salem Ali
Summary: The research focuses on how predictive analytics from machine learning can improve the accuracy of heart disease prediction, with a central focus on minimizing false alarms. Through a feature selection method optimized by dynamic n-gram features, successful supervised learning is achieved to identify the risk of heart disease in potential patients.
Article
Computer Science, Information Systems
Sibren De Bast, Evgenii Vinogradov, Sofie Pollin
Summary: Massive Multiple-Input Multiple-Output (MaMIMO) communication networks have great potential for localization services. However, challenges exist in terms of deploying localized methods in distributed networks and the trade-off between model-driven and data-driven localization. This study proposes two new distributed localization methods and compares them with two centralized methods. The results show that the Machine Learning-enhanced Multiple Signals Classification method (ML-MUSIC) achieves high accuracy, scalability, and reduced fronthaul requirement compared to centralized methods.
Article
Engineering, Electrical & Electronic
Omer Turk
Summary: In this study, a BCI model without any transformation techniques was proposed, aiming to classify cursor down and cursor up movements using EEG signals. The EEG patterns were classified using classic CNN and a hybrid structure of k-NN and SVM with CNN, achieving significantly higher accuracy compared to previous studies in the literature.
INTERNATIONAL JOURNAL OF IMAGING SYSTEMS AND TECHNOLOGY
(2021)
Article
Engineering, Biomedical
K. Mohanavelu, S. Poonguzhali, A. Janani, S. Vinutha
Summary: In order to classify the Pilot's Cognitive Workload (PCWL) level during flight operations, a study conducted in a flight simulator environment used physiological signals and classification algorithms. The study found that HRV features were better for takeoff and landing phases, while EEG features were better for the cruise phase. Using feature selection methods led to improved classification accuracy compared to using all features. LDA and SVM classifiers were more consistent than the kNN classifier in this study.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2022)
Article
Computer Science, Information Systems
Amanpreet Kaur, Munish Kumar, M. K. Jindal
Summary: Image processing is an important research area in computer vision, focusing on recognizing and labeling images based on extracted features. This article specifically explores the use of muzzle point features for cattle identification and compares different feature descriptors and classifiers. The experimental results show that a hybrid feature technique composed of SIFT, SURF, and ORB achieves an accuracy of 97.23% on a dataset of four cattle breeds.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Soobia Saeed, Afnizanfaizal Abdullah, N. Z. Jhanjhi, Mehmood Naqvi, Mehedi Masud, Mohammed A. AlZain
Summary: This study aims to address the challenges of detecting low-grade tumors and CSF in MRI images and proposes a new framework based on a hybrid model of GCSVM and HMMkC. The proposed model shows better results compared to existing techniques and has the potential to assist in identifying brain diseases for patients and doctors.
CMC-COMPUTERS MATERIALS & CONTINUA
(2022)
Review
Computer Science, Software Engineering
Oguzhan Mehmet Gunes, Pelin Kasap, Burcin Seyda Corba Zorlu
Summary: Liver cirrhosis is a significant cause of death globally, emphasizing the importance of early diagnosis. In this study, various machine learning algorithms were employed to classify factors affecting liver cirrhosis, and their performances were compared using real data from 2000 patients. Additionally, the study reviewed over 50 articles on liver disease and classification methods to present the latest developments.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Fernando J. Aranda, Felipe Parralejo, Fernando J. Alvarez, Jose A. Paredes
Summary: Indoor positioning with smartphone-compatible technologies, specifically Bluetooth Low Energy (BLE), has attracted much research attention. This study thoroughly analyzes the state of the art on Wi-Fi and BLE algorithms used for fingerprinting systems and proposes a novel classification scheme. A performance comparison between BLE databases is conducted, evaluating training time, parameter optimization, computational time, and positioning accuracy. The results show that weighted k-nearest neighbors (Wk-NN) algorithms perform better for large scale deployments, while support vector machines (SVM) provide the best accuracy in small environments with less computational and training time.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Jaehong Kim, Jingon Joung, Kiho Lim
Summary: This study examines an intelligent reflecting surface (IRS)-aided space-time line code (STLC) system in order to efficiently extend its coverage area. The phase-shift values of each element of the IRS can be designed to maximize the system's SNR using a semidefinite relaxation (SDR) method. A unit-modulus constraint relaxation (UCR) method is proposed to reduce computational complexity. Numerical results show that the proposed UCR-based IRS achieves comparable bit-error-rate performance to the SDR-based IRS, while significantly reducing computational complexity. The proposed IRS greatly extends the coverage area.
IEEE WIRELESS COMMUNICATIONS LETTERS
(2022)
Article
Engineering, Electrical & Electronic
Ki-Hun Lee, Jingon Joung, Bang Chul Jung
Summary: This letter proposes a novel multiple-input multiple-output transceiver technique, named space-time line coded spatial modulation (STLC-SM). In this technique, two modulation symbols are transmitted through one of the multiple transmit antennas determined by the incoming information bits after constellation-rotated STLC encoding. The receiver decodes the transmitted modulation symbols and the activated antenna index through an optimal joint maximum-likelihood detector. The main result of this study is the mathematically analyzed closed-form upper-bound of bit-error-rate of the proposed STLC-SM system in general antenna configurations. It is validated that the STLC-SM significantly outperforms the conventional STLC in terms of the bit-error-rate.
ELECTRONICS LETTERS
(2022)
Article
Computer Science, Information Systems
Jihoon Choi, Jingon Joung, Yong-Soo Cho
Summary: The study focuses on secure downlink MIMO communications in the presence of a passive eavesdropper, utilizing STLC and AN to enhance physical layer security. The proposed method enables noncoherent detection at the receiver without CSI and achieves full spatial diversity.
IEEE SYSTEMS JOURNAL
(2022)
Article
Computer Science, Information Systems
Seung-Chan Lim, Jingon Joung
Summary: In this paper, a new space-time line code is proposed for multiple-input multiple-output systems with four receive antennas. The paper introduces an encoding method for achieving full rate and a simple decoding scheme that does not require full channel state information at the receiver. Through the assumption of a large number of transmit antennas and analysis of bit error rate performance, it is demonstrated that the proposed line code can achieve full spatial diversity.
IEEE WIRELESS COMMUNICATIONS LETTERS
(2022)
Article
Computer Science, Information Systems
Jingon Joung, Eui-Rim Jeong, Hyukjun Oh
Summary: In this study, a blind symbol and carrier synchronization method is proposed based on the structure of the space-time line code (STLC) signals, taking into account the carrier frequency offset (CFO). The proposed method achieves almost perfect synchronization performance with more than three transmit antennas, regardless of the CFO values. The results indicate that the proposed method is expected to accelerate the practical deployment and operation of low-complexity STLC systems, which achieve full spatial diversity gain with partial or no channel state information at the receiver.
Article
Computer Science, Information Systems
Janghyuk Youn, Jeong Seon Yeom, Jingon Joung, Bang Chul Jung
Summary: We propose a novel cooperative space-time line code (C-STLC) scheme to improve the performance of a relay-assisted internet of things (RIoT). Each relay IoT device (RID) attempts to decode the received signal and sends the STLC encoded signal to an access point (AP) based on the local channel state information (CSI). The AP decodes the received signals without CSI and the proposed C-STLC scheme significantly outperforms conventional cooperative relaying techniques in terms of the outage probability (OP).
Article
Engineering, Electrical & Electronic
Jingon Joung, Jiancun Fan
Summary: This article proposes a linear precoding weight vector to enhance the minimum SINR of multicast users based on STLC. By deriving the optimal weight vector for each user and combining them linearly, the designed linear precoding vector improves the minimum and average SINR of multicast users. Numerical results verify the potential of STLC-based multicast, especially in scenarios with a large number of transmit antennas and severe additive noises.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2022)
Article
Computer Science, Information Systems
Duckdong Hwang, Janghoon Yang, Sung Sik Nam, Jingon Joung, Hyoung-Kyu Song
Summary: We investigate the beamforming for cooperative non-orthogonal multiple access (NOMA) transmission with decode-and-forward relay in full-duplex (FD) or half-duplex (HD) mode. The optimal beamforming for the second-hop of the NOMA transmission is obtained, leading to derived beamforming algorithms for the two relay modes. For FD case, a 2-D search algorithm based on semi definite relaxation (SDR) is provided, where self interference removal is important. For HD mode, no iteration is necessary and only one implementation of SDR is required. Simulation results show the performance advantage of the proposed FD mode system over HD mode in various parameter settings.
IEEE WIRELESS COMMUNICATIONS LETTERS
(2023)
Article
Engineering, Electrical & Electronic
Manzoor Elahi, Jingon Joung, Sungjoon Lim
Summary: The research focuses on using decoupling structure for improving isolation and bandwidth in a MIMO dielectric resonator antenna. The design successfully achieves circular polarization without occupying additional space, resulting in improved bandwidth and isolation performance. Experimental results show overlapping 3 dB axial ratio bandwidth of 13.8% with 20 dB isolation across the bandwidth.
IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS
(2023)
Article
Telecommunications
Jingon Joung, Heejung Yu, Tony Q. S. Quek
Summary: This study focuses on addressing the challenges of interference in integrated sensing and communication (ISAC) systems and highlighting the advantages of multi-functional ISAC schemes. The authors propose a zero-forcing (ZF)-based beamforming and power allocation strategy and emphasize the importance of adaptive balance between sensing and communication capabilities. The research demonstrates that relying solely on single function-aware schemes is insufficient and that integrated sensing and communication functionalities are crucial for efficient ISAC systems.
VEHICULAR COMMUNICATIONS
(2023)
Article
Computer Science, Information Systems
Hunje Lee, Sujeong Han, Jeong-Il Byeon, Seoulgyu Han, Rangun Myung, Jingon Joung, Jihoon Choi
Summary: This paper proposes a method for detecting and classifying drones using sensor fusion schemes. The method combines optical images, radar range-Doppler maps, and audio spectrograms using convolutional neural networks (CNN) and multinomial logistic regression. Numerical simulations show that the proposed sensor fusion method can significantly improve the detection and classification accuracy of drones compared to individual sensing schemes.
Article
Computer Science, Information Systems
Kyubin Kim, Jaehong Kim, Han-Gyeol Lee, Jihoon Choi, Jiancun Fan, Jingon Joung
Summary: This paper proposes a vision-based unmanned aerial vehicle (UAV) chasing system that can protect a pursuer UAV (pUAV) from attacks by an evader UAV (eUAV). The system consists of UAV tracking and control signal generation, using a combination of YOLOv3 and object trackers to improve tracking performance. The control signals are generated based on the predicted bounding box area and proportional-derivative control method. Evaluation results show that the proposed algorithm outperforms YOLOv3 in terms of object tracking accuracy and computational complexity reduction.
Article
Computer Science, Information Systems
Jingon Joung, Soobum Park, Ji-Myung Oh, Eui-Rim Jeong
Summary: This study examines a cooperative communication system using multiple decode-and-forward relay nodes (RNs) that perform phase rotation of the regenerated signals before retransmitting them to a destination node (DN). A received signal-to-noise ratio (SNR) threshold-based RN association method is proposed for communication from a source node to RNs. The study also explores the optimal number of phase rotations (PRs) in a frame and shows that more PRs provide better bit-error-rate (BER) performance at the DN.
Article
Computer Science, Information Systems
Subin Kim, Han-Gyeol Lee, Songmin Lee, Jaehong Kim, Jingon Joung, Juyeop Kim
Summary: This paper demonstrates the implementation of the SFLC-OFDM system and proposes frequency-domain signal processing methods and optimization methods for real-time operating systems.
Article
Engineering, Electrical & Electronic
Ki-Hun Lee, Jeong Seon Yeom, Jingon Joung, Bang Chul Jung
Summary: This study investigates uplink non-orthogonal multiple access (NOMA)-based IoT networks using constellation-rotated space-time line code (CR-STLC). The authors mathematically analyze the bit-error-rate (BER) performance and spatial diversity order, showing that the system achieves optimal diversity order in the high signal-to-noise ratio (SNR) regime. Two different rotation angle optimization techniques are considered, with the fixed rotation approach yielding similar performance as the dynamic approach. The study also demonstrates that the CR-STLC NOMA system achieves spatial diversity order of 1 even with simultaneous transmission from multiple STAs, providing improved spatial diversity gain compared to systems without constellation rotation.
IEEE OPEN JOURNAL OF THE COMMUNICATIONS SOCIETY
(2022)