Yazar "Salah, Bashir" seçeneğine göre listele

Listeleniyor 1 - 4 / 4
  • Küçük Resim Yok
    Öğe
    Acoustic signal-based indigenous real-time rainfall monitoring system for sustainable environment
    (Elsevier, 2023) Kumari, Rani; Sah, Dinesh Kumar; Cengiz, Korhan; Ivkovic, Nikola; Gehlot, Anita; Salah, Bashir
    The rainfall weather station employs a tipping bucket rain gauge, which serves as a specialized instrument for the meticulous assessment and documentation of various rainwater parameters. The implementation of a tipping bucket rain gauge for rainfall monitoring bears significant implications for both societal productivity as well as improvement of human life. A noteworthy example can be the constructive influence of rainwater over the sustainable agricultural irrigation practices, wherein the precise monitoring of rainfall through a tipping bucket rain gauge enables the formulation of tedious irrigation strategies. The rainfall monitoring if often handle using rain gauge which majorly faces two challenges named as mechanical devices failure and high installation and maintenance cost. Considering the challenges, we propose the fully automated rain gauge (RG) based on the principle of sound and its properties for rainfall monitoring. The working prototype is part of our work whose primary task is to collect the rainfall acoustic value and store it in the cloud. Our mechanism is to use the acoustic property of rain data to categorize rainfall intensity. We perform blind signal separation on the received signal (acoustic signal recorded with the help of microphone sensor) and feed the separated signal to a recurrent convolution neural network (RCNN). The source separation of the collected acoustic signals is primarily being done using independent component analysis and principal components analysis. The proposed solution can be able to make the classification of rain intensity with more than 80% accuracy. In addition to this, the developed method provides the sustainable solution to the challenges with the low-cost and application-specific acceptable threshold criteria and supplement rain measurement techniques.
  • Küçük Resim Yok
    Öğe
    BukaGini: A Stability-Aware Gini Index Feature Selection Algorithm for Robust Model Performance
    (Ieee-Inst Electrical Electronics Engineers Inc, 2023) Bouke, Mohamed Aly; Abdullah, Azizol; Frnda, Jaroslav; Cengiz, Korhan; Salah, Bashir
    Feature interaction is a vital aspect of Machine Learning (ML) algorithms, and gaining a deep understanding of these interactions can significantly enhance model performance. This paper introduces the BukaGini algorithm, an innovative and robust approach for feature interaction analysis that capitalizes on the Gini impurity index. By exploiting the unique properties of the BukaGini index, our proposed algorithm effectively captures both linear and nonlinear feature interactions, providing a richer and more comprehensive representation of the underlying data. We thoroughly evaluate the BukaGini algorithm against traditional Gini index-based methods on various real-world datasets. These datasets include the High School Students' Performance (HSSP) dataset, which examines factors affecting student performance; Cancer Data, which focuses on identifying cancer types based on gene expression; Spambase, which targets spam email classification; and the UNSW-NB15 dataset, which addresses network intrusion detection. Our experimental results demonstrate that the BukaGini algorithm consistently outperforms traditional Gini index-based methods in terms of accuracy. Across the tested datasets, the BukaGini algorithm achieves improvements ranging from 0.32% to 2.50%, underscoring its effectiveness in handling diverse data types and problem domains. This performance gain highlights the potential of the BukaGini algorithm as a valuable tool for feature interaction analysis in various ML applications.
  • Küçük Resim Yok
    Öğe
    Integrating CLDs and machine learning through hybridization for human-centric wireless networks
    (Wiley, 2024) Kumari, Binita; Yadav, Ajay Kumar; Cengiz, Korhan; Salah, Bashir
    Wireless sensor networks, more commonly abbreviated as WSNs, have been regarded as helpful tool for managing human-centric applications. Nevertheless, the design of wireless systems that are accurate, efficient, and robust remains difficult due to the variables and dynamics of the wireless environment as well as the requirements of the users. Cross-layer designs along with the machine-learning techniques need to be integrated into a novel hybridization framework for human-centric wireless networks in order to simplify the process and make it more manageable. The purpose of the proposed framework is to enhance wireless sensor networks (WSNs) in terms of their energy efficiency, robustness, real-time performance, and scalability. In particular, machine learning are employed for the purpose of extracting features from sensor data, and the framework combines cross-layer optimization and RL in order to facilitate effective and adaptable communication and networking. In comparison to previous work in this field, the accuracy, energy consumption, robustness, real-time performance, and scalability of the proposed framework are all significantly improved. The hybridization framework that has been proposed provides a promising approach to addressing the challenges, and it can be of use to a variety of applications. A diagram of the proposed hybridization framework. The wireless sensor network (WSNs) collects sensor data and uses a convolutional neural network to extract features. The features are then used by a machine learning algorithm to generate control signals for the system. The control signals are then used to adjust the behavior of the wireless network, which in turn sends configuration commands back to the sensor network.image
  • Küçük Resim Yok
    Öğe
    SOHCL-RDT: A self-organized hybrid cross-layer design for reliable data transmission in wireless network
    (Elsevier, 2023) Cengiz, Korhan; Kumari, Rani; Sah, Dinesh Kumar; Ivkovic, Nikola; Salah, Bashir
    In this paper, we propose SOHCL-RDT'' which stands for a self-organized hybrid cross-layer design for reliable data transmission in wireless network. The communication paradigm is changing and new approach related to machine learning or other optimization algorithms are being introduce rapidly. The TCP/IP or OSI model is not at all equipped to accommodate such a vast changes in its established protocol stacks. Considering this, we have proposed the hybrid cross layer design where the communication or transmission will be handle using two set of protocol stack. One set for the established classical network, and another using cross layer approach. Our design leverages the strengths of both the physical and MAC layers to optimize packet transmission and minimize energy consumption. An optimization algorithm based on gradient descent is also developed to adjust transmission parameters in real-time. The objective is to invoke the classical model only when it needed; it means until unless gradient descent is able to make classification regarding the node scheduling and achieve the acknowledgment, the TCP/IP protocol stack will be in deactivation. Using this method, we have performed our experiments mainly on two parameters named as packet delivery ratio (PDR), end-to-end delay (E2ED); because these are important aspect of reliability. In addition to that, the energy consumption of network is also compared with the existing algorithms. The results show that the proposed hybrid cross-layer design outperforms the existing algorithms. The performance gain can be attributed to the cross-layer approach and the use of the optimization algorithm. Overall, the proposed hybrid cross-layer design is a promising solution for reliable data transmission in wireless sensor networks, with the potential to improve network performance and prolong network lifetime by reducing energy consumption.& COPY; 2023 Published by Elsevier B.V.