Archives
Volume-1, Issue-1, July - Dec, 2024
Abstract: Diabetes is a growing global health concern, with its rates increasing across all age groups, affecting children, teenagers, adults, and older individuals. This study explores the prediction of diabetes using machine learning algorithms, including K-Nearest Neighbors (KNN), AdaBoost, and Random Forest, with a dataset from the National Institute of Diabetes and Digestive and Kidney Diseases. The dataset includes diagnostic measurements like the number of pregnancies, glucose levels, blood pressure, skin thickness, insulin, body mass index (BMI), diabetes pedigree function, and age, with an outcome variable indicating whether diabetes is present. Among the algorithms tested, Random Forest achieved the highest accuracy at 74.03%, compared to 66.23% for KNN and 73.38% for AdaBoost. Based on these results, a predictive model was created using Random Forest, along with a userfriendly interface built with the Streamlit Python library. This interface allows users to register, log in, and fill out a medical form to assess their risk of diabetes. By clicking the "Test Results" button, users receive predictions indicating whether they have diabetes based on their input data. User authentication is managed through MongoDB, ensuring secure storage and validation of credentials. The application is deployed on Render, providing easy access for users on both mobile devices and laptops, making it a convenient tool for evaluating diabetes risk based on personal health information.
Abstract:The use of wireless sensor networks (WSN) has increased recently with its various and real-time applications, which make them vulnerable for various types of attacks, considering their limited resources we need an automatic system which will be able to detect, alert and react as fast as possible against any abnormal behaviors. The traditional protocols of security like cryptography, key management, routing protocols is no longer useful. A system based on artificial intelligence, specifically learning models present additional value and can help to increase security in wireless networks performed by their high accuracy and provide a good model for reducing the computational cost. In this paper, we provide an overview on wireless sensor networks; we highlight their vulnerabilities, constraints and the current solutions of security.
Abstract:This review article discusses several deep learning methods, such as generative adversarial networks (GANs), recurrent neural networks (RNNs), and convolutional neural networks (CNNs), that are utilised in surgical image processing. The strengths and limitations of each technique are discussed, including accuracy, computational efficiency, and the ability to handle complex data. The comparison parameters for evaluating the performance of these techniques are accuracy, speed, and scalability. The applications of deep learning in surgical image analysis, such as preoperative planning, intraoperative guidance, and postoperative analysis, are also covered. This review highlights the importance of considering these comparison parameters when choosing a deep learning technique for a specific surgical image analysis task. Deep learning has become a powerful tool in surgical image analysis, with a range of techniques available to choose from. The choice of technique will depend on the specific task and dataset being analyzed, as well as the comparison parameters that are most relevant to the project at hand. With the advancement of deep learning, we can expect to see even more sophisticated techniques being developed and applied in surgical image analysis.
Abstract:This study investigates the effect of biopolymers, specifically Xanthan gum, on the geotechnical properties of red soils, a critical aspect of soil stabilization in civil engineering. Traditional stabilization methods often rely on chemical additives, which pose environmental concerns. Biopolymers, with their inherent biodegradability and low environmental impact, have emerged as a promising eco-friendly alternative. The study examines various biopolymers and their impact on the shear strength, compaction, and water retention of red soils, with a focus on optimizing biopolymer concentrations. Laboratory experiments, including Unconfined Compressive Strength (UCS), Proctor, and California Bearing Ratio (CBR) tests, were conducted to evaluate the mechanical behavior of soil with biopolymer additives. The results show a significant improvement in soil strength and stability, suggesting biopolymers as a sustainable solution for soil stabilization. This study contributes to the growing body of research on biopolymer-based soil stabilization, highlighting their potential benefits and real-world application in infrastructure projects.
Abstract:Food image categorization is an emerging research area due to its increasing significance in the health and medical domains. The development of automated food recognition techniques holds great promise for applications such as diet monitoring systems and calorie estimation. In this study, we explore automated food classification methods utilizing deep learning algorithms. Specifically, we employ SqueezeNet and VGG-16 Convolutional Neural Networks (CNNs) for food image classification. To enhance the performance of these networks, we apply data augmentation techniques and fine-tune the hyperparameters. These optimizations result in improved accuracy, rendering the networks suitable for practical applications in the health and medical fields. SqueezeNet, known for its lightweight structure and ease of maintenance, achieves high accuracy even with fewer parameters. By extracting complex features from food photographs, the accuracy of food image classification is further enhanced. Additionally, our proposed VGG-16 network demonstrates notable advancements through increased network depth. Overall, our research highlights the efficacy of deep learning-based automated food image classification. The findings underscore the potential of these techniques in revolutionizing various domains, including health, medicine, and dietary analysis.