International Journal of Communication Networks and Information Security (IJCNIS)
Not a member yet
1021 research outputs found
Sort by
Effective Feature Selection Mechanism in Semi-Supervised Sentiment Analysis on E-Commerce Reviews
No full textReceived: 27Apr 2024
Accepted: 04 Sep 2024
In this globalized world, people prefer to buy products online without any hesitation. Usually, to acquire the quality of the product or brand, they examine the product’s reviews, which is a tedious job to do manually. The wide use of social media also encourages the users, to keep their views on the product in a global platform. By using machine learning techniques, we can solve the problem of product selection. In this study, we are using Sentiment Analysis to analyze the reviews and select the best features. We have applied Support vector Machine and Naïve Bayes machine learning algorithms for the binary classification of the reviews, where it tells whether the review is favorable or not, i.e. Positive or Negative. The problem with the real-time review analysis is that all the reviews we are considering for the analysis are not labeled. So, we are using a semi-supervised machine learning technique to retrieve the missing information from the E-commerce product reviews for better information and improved accuracy. Additionally, we want to address the issue of sentiment polarity categorization, boost productivity, and gain a deeper understanding of how sentiment analysis may be used to inform business decisions. As a result, this research can help consumers understand the knowledge of product reviews and justify the product quality based on the data i.e., reviews. This study is carried out with two popular semi-supervised methods, self-training and co-training, and implemented on the E-Commerce dataset. As a result, it found that the Co-Training model with support vector machine and Naïve Bayes classifiers performs better than the Self-Training model with support vector machine classifier for the dataset which contains both the labeled and unlabeled data
Gradient Directional Edge Coding (GDEC) for Expression Recognition from Facial Images
No full textFacial Expression Recognition (FER) requires an effective Expression descriptor that can provide sufficient discrimination between different facial expressions. However, the existing FER methods are susceptible to noise, distortions and not able discriminate flat regions from noisy regions. Hence, this paper proposes a new Face descriptor called as Gradient Directional Edge Coding (GDEC) which encodes the expression components through the directional edges. Initially, GEDC finds the gradients for each pixel and then encodes them with their neighbor pixel’s support. The support is assessed based on the deviation in the direction of corresponding neighbor pixel with the mean direction of a local region. Each pixel is encoded a 7-bit code word among which the six bits are belongs to the directions of neighbor pixels and one bit is sign bit. After describing the expression, the classification is accomplished through Support Vector Machine at different kernels. Experimental validation on Standard CK+ dataset shows an accuracy of 94.6300% which is outstanding compared to the state-of-the-art methods
MEASURING THE IMPACT OF SOCIAL NETWORKS ON WOMEN ENTREPRENEURS' SUCCESS: AN AHP ANALYSIS
No full textThis study investigates the impact of various social networks on the success of women entrepreneurs using the Analytic Hierarchy Process (AHP). Recognizing that social networks play a crucial role in providing resources, emotional support, and business opportunities, this research evaluates three types of networks: Local Community Networks, Professional Associations, and Online Networks. The study finds that Local Community Networks offer the highest levels of Social Capital and Emotional Support, which are critical for overcoming the unique challenges faced by women entrepreneurs. Professional Associations provide valuable structured resources and mentorship, while Online Networks, despite their broad reach, often offer weaker ties. The AHP model proves effective in prioritizing these networks based on their contribution to entrepreneurial success, highlighting the importance of a balanced approach that combines strong local and professional connections with strategic online engagement
Evaluating Open Data through a Confidence Assessment Matrix Model: Balancing Information Security, Societal, and Economic Impact
No full textThis paper presents the Confidence Assessment Matrix Model (CAMM), an innovative framework aimed at evaluating information security, social, and economic impacts of open data. The study addresses the challenge of assessing the reliability and impact of open data, which are often used without adequate evaluation of their quality and influence. The primary objective of CAMM is to measure confidence levels in these portals by integrating expert judgments and user familiarity with datasets. CAMM incorporates various criteria such as information security, data privacy, accessibility, usability, and the potential socio-economic effects of the datasets. Using the Dempster-Shafer Theory (DST), the model provides a structured method to handle uncertainty and offers a quantitative assessment of the likelihood and confidence associated with different risk and benefit factors. CAMM operates through three key layers: (1) Expert Judgment Quantification, which captures expert assessments; (2) Risk and Benefit Trade-off Matrix, which evaluates the balance between potential impacts, including security vulnerabilities; and (3) New Confidence Level, offering a synthesized measure of reliability and safety. The results demonstrate CAMM’s ability to effectively assess the social and economic implications of open data, highlighting its significance in guiding data-driven decision-making. The findings suggest that CAMM can serve as a valuable tool for policymakers and stakeholders in optimizing the use of open data for societal benefit while safequarding information security.
 
An Extensive Review of Developments and Methods in Super-Resolution Image Reconstruction
No full textSuper-resolution (SR) image reconstruction plays a vital role in enhancing the resolution of low-resolution (LR) images, benefiting various fields such as remote sensing, medical imaging, and surveillance. The SR problem entails reconstructing high-resolution (HR) images from LR inputs, typically due to the loss of high-frequency information and the problem's ill-posed nature. This review study presents a comprehensive overview of recent developments and approaches in SR image reconstruction. This study primarily presents three approaches: methods based on learning, methods based on reconstruction, and methods based on interpolation. Despite the fact that interpolation-based approaches, such as bicubic interpolation, are straightforward and quick, they frequently result in blurring and loss of high-frequency features. Reconstruction-based methods leverage prior knowledge of image characteristics to recover HR images, often through optimization techniques. However, these methods may suffer from slow convergence and high computational cost. Because of their capacity to learn complicated mappings between LR and HR picture spaces, learning-based methods—and deep learning approaches in particular—have been the center of a lot of attention lately. These methods leverage large datasets to train convolutional neural networks (CNNs) for image super-resolution, achieving remarkable performance in terms of visual quality and computational efficiency. Furthermore, we discuss the challenges and future directions in SR research, including the development of more robust and efficient algorithms, handling noisy real-world data, and exploring novel architectures and loss functions to further improve SR performance. The purpose of this review paper is to provide a comprehensive overview of strategies for SR image reconstruction. It focuses on the progression from conventional interpolation methods to cutting-edge deep learning approaches. They hope that this publication will serve as a valuable resource for scholars and practitioners in the field of computer vision and image processing
Dynamic Sweet Spot Audio System: Enhancing Personalized Audio Signal Through ILD and ITD Techniques
No full textThe proposed model aims to create a audio system that can adjust the optimal listening area, known as the "sweet spot," to the listener’s specific location. The sweet spot is the ideal position where the sound is most balanced and clear. Traditional audio systems have a fixed sweet spot where the audio is heard as intended by the mixer. This new system, however, allows for the adjustment of the sweet spot so listeners can have a personalized listening experience with high-quality sound tailored to their preferences. Instead of moving the speakers, this model directs the audio signals from the speakers towards the listener’s position. This approach is innovative, as no existing product on the market offers this capability. The system utilizes ILD (interaural level difference) and ITD (interaural time difference) methodologies to tailor the audio signals to the listener’s specific needs
IoT-Based Biometric Attendance System Using Arduino and ThingsBoard
Get PDFThe fingerprint sensor is a sensor that detects fingerprints using an optical system, where detection is done by reading the contours of the fingerprints and the static electricity of the body. However, data generated by fingerprint sensors, in general, can only be accessed if connected directly to the fingerprint module. From these conditions, operational managers or business managers can't monitor the absence of discipline of their employees because attendance data cannot be accessed directly. They must go through the download process from the machine's default software. The purpose of this research is to build a fingerprint recognition system in the context of an abscess machine on an IoT basis so that fingerprint data processing is centralized so that it can be easily accessed without having to connect directly with the fingerprint sensor module that is available by implementing the client-server method. The test results of this study indicate that collaboration between the fingerprint sensor module integrated with the Arduino Uno module and the ThingsBoard IoT platform can be done with a fingerprint reading accuracy of 96.25%, and data can be accessed in real time through the ThingsBoard server
Adaptive Scheduling Algorithm for Live Video Streaming in P2P Network
Get PDFThe goal of this paper is to develop a network code that will help improve the performance of live streaming systems. The major factors that influence the network bandwidth is the buffer map updating delay between the peers, The proposed model deals with the scheduling algorithm which identifies the required number of packets to the various parent nodes. The scheduling algorithm also considers the encoding and forwarding mechanisms in the network. The proposed method improves the streaming continuity in the network and reduces redundancy. The experimental results claim that the proposed algorithm improves the video quality and minimizes the rate of redundancy. It also provides a high packet rate compared to the existing algorithms
Enhancement of Buffer Management and Data Transmission in Delay Tolerant Network Using Secant Osprey Optimization
Get PDFIn both wired and wireless networks, message forwarding takes place after establishing a route between source and destination, even in the infrastructure-less network. It stores the messages in buffer nodes and carries them until a destination or appropriate relay node is encountered. However, the storing and carrying approach for a long time causes buffer congestion. Therefore, an optimized technique is required for buffer management and efficient communication. In Delay Tolerant Network (DTN), the chances of disconnections are higher due to propagation issues, node mobility, and power disruptions. In order to solve the problem, this paper proposed a Secant Osprey Optimization (SOO)-based efficient buffer management scheme. First, the message data is collected from the Twitter dataset for buffer analysis, and then the repeated messages are removed using SHA-512. Then, messages are clustered using Lorentz Distributed K-Means (LDKM). Further, the messages are scheduled using a Probabilistic Function-based Adaptive Neuro Fuzzy Inference System (PF-ANFIS) for forwarding. For forwarding, the optimal path is selected using the SOO algorithm. Thus, the heavy load data are forwarded via the selected path without disruptions. The experimental analysis is carried out using the PYTHON software tool by comparing the proposed model with the existing methods. The simulated outcome showed that the proposed methodology attains a higher delivery ratio of 0.815 with a lower delay rate of 1217s. Also, the proposed technique schedules the messages with 98.2% accuracy, 96% precision, and 96.5% recall. Hence, the presented approaches are more highly performed in the buffer management of DTN than the existing techniques
Analysis of CMOS IC-based Hybrid Architecture for Edge Computing
Get PDFWith the rapid advancement of Internet of Things (IoT), mobile internet, and big data technologies, edge computing has emerged as a novel computing paradigm. In the hybrid architecture of edge computing,Complementary Metal-Oxide-Semiconductor (CMOS) integrated circuits play a pivotal role in empowering edge devices and servers with essential computing, storage, and communication capabilities. Despite their critical importance, CMOS integrated circuits in edge computing environments confront significant challenges in low-power electronics. These challenges include an increase in power density and a decrease in system stability and reliability. This paper delves into the key technologies of the hybrid architecture in edge computing and sheds light on the vital role of CMOS integrated circuits in edge devices. It introduces a novel approach for low-power electronics, which encompasses methods like optimization of double threshold voltage and refinement of algorithmic processes. These methods aim to tackle the power-efficiency issues while maintaining the performance of edge computing systems.Furthermore, the paper presents a detailed analysis of the proposed low-power techniques, focusing on how they can effectively reduce power consumption without compromising the functionality and efficiency of the edge computing systems. It concludes with a comprehensive discussion on the optimization results, highlighting the benefits and potential implications of implementing these low-power strategies in edge computing environments. This discussion not only underscores the importance of energy efficiency in edge computing but also opens new avenues for future research and development in this rapidly evolving field