Volume 8, Issue 1 - IJIRTM
January - February (2024)
Impact Factor: 5.86 | Volume 10 | Issue 4
A Study to Assess the Effectiveness of Planned Teaching Program Regarding Stem Cell Preservation among Antenatal Mothers in Rural and Urban Area of Bhopal City Madhya Pradesh
👥 Dushyant Sharma
📙 Abstract : Mother and baby share a perfect bond from the period of conception and it is she who nurtures and gives the best of everything to her child. And with the advancement of the technologies she is not just bound to care for her baby at the present but she can now gift her baby with a gift of health, through stem cell and cord blood banking this is effort of science for construction of technology for tomorrow. To assess the pre exiting knowledge of antenatal mother regarding stem cell preservation. To find out the effectiveness of Planned Teaching Programme regarding antenatal mother regarding stem cell preservation by comparing pretest and posttest knowledge score. To find out the associations between the pretest knowledge score with selected demographic variables.
🔖 Keywords :️ Stem Cell Preservation, Antenatal, Mothers.
Congestion-Aware Multi-Route Establishment Routing for Mobile Ad Hoc Networks (MANET)
👥 Kanchan Narware, Chetan Agrawal, Pooja Meena
📙 Abstract : This research delves into the realm of Mobile Ad-hoc Networks (MANETs), dynamic and infrastructure-less wireless communication systems that offer flexibility for deployment at any time and place. With the nodes in MANETs constrained by limited battery capacity, energy optimization becomes a pivotal design consideration. This paper comprehensively reviews prior works aimed at enhancing the longevity and various performance parameters of MANETs. A primary challenge in MANETs is congestion, arising from the restricted link capacity between nodes. To tackle this issue, the research proposes a resolution using the Ad-hoc On-demand Multipath Distance Vector (AOMDV) routing protocol, introducing alternative paths to alleviate congestion. Comparative analysis with the MEALBM scheme underscores the superior performance of the proposed approach, affirming its effectiveness in addressing congestion and contributing to the overall enhancement of MANET capabilities.
🔖 Keywords :️ MANET, Queue, Load balancing, Routing, Multipath, MEALBM, Congestion.
A Review on Cloud Computing, Security Issues and Techniques
👥 Dinesh Kumar Malviya, Pooja Meena, Chetan Agrawal
📙 Abstract : Cloud computing is a network-built invention that allows users to access information whenever they need it. Existing cloud computing systems have stringent safeguards in place to preserve user data confidentiality. Cloud computing is made up of a number of technologies and regulations to protect infrastructure, services, and data. Since the infrastructure is not owned by the customer, the conventional security architecture is quite difficult to implement. The paper presented the review associated with cloud computing as well as strategies for users to mitigate these risks and problems. The paper also covers the current challenges for cloud computing to protect their infrastructure from threats and hackers.
🔖 Keywords :️ Cloud Computing, Security Issues, Techniques Used, Encryption.
No-Shows Appointment Prediction Using Machine Learning: A Review
👥 Vaishali Shukla, Pooja Meena, Chetan Agrawal
📙 Abstract : Clinics and hospitals globally are losing substantial profits due to patient no-shows. This challenge is faced by any healthcare facility allowing advance appointment bookings, as patients may not show up or cancel too late. To mitigate this, many hospitals use costly reminder systems and overbooking strategies, where multiple patients are booked for the same slot. These methods, however, do not fully address the economic and social impacts of high no-show rates. To tackle this issue, the paper presented a review on techniques used for appointment scheduling and intuitive management systems that includes features for automated handling of high-risk appointments using case-specific outcome predictions, thereby aiming to reduce idle time and overtime for medical staff. The paper also shows the role of machine learning for maintaining efficient handling of such system.
🔖 Keywords :️ No-Shows, Missed-Appointment, Machine Learning.
Examining Diabetic Retinopathy (DR) Through Deep Learning
👥 Gagan Deep Kaur, Pooja Meena, Chetan Agrawal
📙 Abstract : Diabetic retinopathy (DR) is a complex ailment affecting individuals with diabetes, leading to retinal damage and the potential onset of blindness. This condition disrupts retinal blood vessels, resulting in fluid leakage and severe distortion of vision. DR is a prevalent eye disease linked to chronic diabetes and stands as the primary cause of blindness among working-age adults globally, potentially impacting more than 93 million individuals. This study introduces an automated classification system designed to analyze fundus images under varying illumination and fields of view. Leveraging machine learning models such as Otsu, Random Forest, and Clehe algorithm, the system generates severity grades for diabetic retinopathy. The incorporation of machine learning models, particularly Random Forest, brings forth the advantage of high variance and low bias. This characteristic allows the classifier not only to diagnose diabetic retinopathy but potentially detect a broader spectrum of nondiabetic diseases. Visualizations of features learned by the Region-Based Convolutional Neural Network(RCNN) and Gray Level Matrix Co-occurrence (GLMC) demonstrate that the signals used for classification predominantly originate from clearly observable parts of the image. Notably, moderate and severe diabetic retinal images showcase macroscopic features aligned with the architecture's training and validation accuracy. This research introduces a promising avenue for automated diabetic retinopathy severity classification, offering potential advantages in early diagnosis and intervention for the eye health of diabetic patients.
🔖 Keywords :️ Diabetic Retinopathy, Classification, Image Processing, Deep Learning, Segmentation, Severity Grade.
Enhancing Text Classification Performance through Machine Learning Techniques
👥 Sheetal Jaiswal, Chetan Agrawal, Rashi Yadav
📙 Abstract : This research paper focuses on leveraging machine learning techniques to enhance text classification performance for marketing-related product reviews. The primary objective is to discern the sentiment of customer reviews, thereby aiding in improving service and product quality. Acquiring valuable online reviews presents a challenge, particularly in accurately recognizing review sentiment. The study dynamically identifies customer opinions regarding specific products, distinguishing between positive and negative sentiments. Given the significance of online reviews in purchase decision-making, this work contributes to monitoring and managing early promotion efforts. Furthermore, since reviewers often translate into actual product adopters, precise sentiment identification facilitates direct purchases. The paper presents a method utilizing Support Vector Machine (SVM) and Naive Bayes algorithms for text classification, demonstrating superior accuracy in Text Classification.
🔖 Keywords :️ Machine Learning, Text Classification, SVM, Sentiment Analysis.
A Detailed Survey of Color Image Encryption Standards (IES)
👥 Mahak Saxena, Kaptan Singh, Amit Saxena
📙 Abstract : Color image encryption is an essential domain within the realm of information security, ensuring the confidentiality and integrity of visual data transmitted over networks or stored in digital form. This paper presents a comprehensive survey of existing color image encryption standards (IES), examining their methodologies, strengths, weaknesses, and applicability across diverse scenarios. By systematically reviewing prominent encryption techniques tailored specifically for color images, this survey aims to provide researchers and practitioners with a structured understanding of the state-of-the-art in color image encryption. Through a detailed analysis of various standards, including their encryption algorithms, key management strategies, and performance metrics, this paper offers insights into the evolving landscape of color image security. Additionally, it discusses emerging trends and potential avenues for future research, thereby facilitating advancements in the development of robust and efficient color image encryption solutions.
🔖 Keywords :️ Image encryption, Decryption, Latin square, Security, Algorithms.
Deployment and Use of Dark net on Private Network: The Future of Intrusion Detection System
👥 B.K. Verma
📙 Abstract : With the development of computer network, network security has become more important issue to the users and organizations. As the use of computer network is increasing, security of the data and information on the network became a major concern. As the importance and use of the computer network increases, rapid identification of threats at a global level becomes even more complex. Better advance warning benefits the entire world of computer network. A Darknet is a portion of network, a certain routed space of IP Addresses in which there are no active servers or services. I.e., externally no packet should be directed to that address space. Most are the systems are designed for the public internet monitoring. Large or mid-sized organization can also take benefits of the traffic entering on darknets to identify the threats coming on their network. In this paper we have described the issue of identifying dark IPs in the network. We have analysed various methods to identify dark IPs the in private network and proposed method than can be implement on private network to identify dark IPs. This is the research work and it is not fully implemented. The implementation of this research work is on-going.
🔖 Keywords :️ Darknet, Dark IPs, Private Network, IDS, Network Security, ARP.
Federated Learning for Privacy-Preserving AI
👥 Gulabpreet Singh, Sabhyata Kalra, Utkarsh, Karan, Soumya Tripathi
📙 Abstract : Modern machine learning models are typically trained by centralizing data on a single server, an approach that conflicts with growing privacy regulation and with the practical reality that much valuable data, patient records, financial transactions, on-device usage logs, cannot or should not leave the organization or device where it was generated. Federated Learning (FL), introduced by McMahan et al. as the Federated Averaging algorithm, addresses this by training a shared model across many decentralized clients, each of which computes an update using its own local data, sending only that update, not the raw data, to a central server for aggregation. This paper reviews the architecture, privacy-enhancing techniques, and application landscape of federated learning. We describe the core FedAvg algorithm and the cross-device versus cross-silo deployment settings, then survey the privacy and security layer built on top of FL, including differential privacy, secure aggregation, and homomorphic encryption, each of which defends against a different threat to the base protocol's privacy guarantees. We then review applications across four domains: healthcare, where FL enables multi-institutional model training without sharing patient records; Internet-of-Things (IoT) and smart devices, where FL enables on-device personalization; and banking and finance, where FL supports collaborative fraud detection and credit risk modeling across institutions that cannot share raw customer data. We close with a discussion of open challenges, including statistical heterogeneity across clients, communication efficiency, robustness to malicious participants, and regulatory alignment, and outline directions for future research.
🔖 Keywords :️ Federated learning, Privacy-preserving AI, Differential privacy, Secure aggregation, Healthcare AI, Internet of Things, Financial fraud detection.