Volume 10, Issue 4 - IJIRTM
July - August (2026)
Impact Factor: 5.86 | Volume 10 | Issue 4
Intrusion Detection System based on Machine Learning and Deep Learning Model
👥 Ankit Gupta, Dr.Harsh Mathur
📙 Abstract : The exponential growth of cloud computing, Internet of Things (IoT), Industrial Internet of Things (IIoT), edge computing, and intelligent networking has significantly increased the frequency and complexity of cyber-attacks. Traditional signature-based Intrusion Detection Systems (IDS) are often ineffective against zero-day attacks, polymorphic malware, and sophisticated network intrusions due to their dependence on predefined attack signatures and limited adaptability. Consequently, Machine Learning (ML) and Deep Learning (DL) techniques have emerged as promising solutions for intelligent intrusion detection by automatically learning network traffic patterns and identifying malicious activities. This literature survey critically reviews recent advances in ML- and DL-based intrusion detection systems, compares existing methodologies using standard performance metrics such as Accuracy, Precision, Recall, F1-score, False Positive Rate (FPR), False Negative Rate (FNR), Area Under the Curve (AUC), and Detection Rate (DR), and identifies current research gaps. This research work presents, comparative study for intrusion detection system using ML-DL based approaches, and also provides valuable insights for designing next-generation intelligent, scalable, explainable, and lightweight intrusion detection systems suitable for modern cloud, IoT, and edge computing environments.
🔖 Keywords :️ Machine learning, Network security, Deep learning, Intrusion detection system, Cyberattacks.
AI-Based Lung Image Segmentation and Compression for Medical Image Analysis
👥 Puneet Tiwari, Dr.Anupam Chouksey
📙 Abstract : Medical image analysis plays an important role in the diagnosis and treatment of lung-related diseases. The increasing availability of high-resolution lung images, including CT scans and chest X-ray images, has created challenges related to accurate image segmentation, storage, processing, and transmission. In this paper, we propose an AI-based lung image segmentation and compression framework for efficient medical image analysis. The proposed approach uses deep learning techniques to automatically identify the lung region from medical images and subsequently compress the extracted region while preserving important diagnostic information. Initially, lung images are collected from publicly available medical image datasets and subjected to preprocessing operations such as resizing, noise removal, normalization, and enhancement. A deep learning-based segmentation model is then used to distinguish the lung region from the surrounding anatomical structures. The segmented lung region is subsequently processed using an autoencoder-based compression technique to generate a compact representation of the medical image. The reconstructed image is evaluated to determine the quality of the compression. The proposed framework can support efficient storage, transmission, and subsequent analysis of medical images. Segmentation performance can be evaluated using Dice coefficient, Intersection over Union, precision, recall, and accuracy, while compression performance can be evaluated using Mean Squared Error, Peak Signal-to-Noise Ratio, Structural Similarity Index, and compression ratio. The proposed AI-based framework provides an integrated solution for automated lung segmentation and efficient medical image compression and can serve as a preprocessing stage for computer-aided diagnosis systems.
🔖 Keywords :️ Artificial Intelligence, Lung Image, Medical Image Analysis, Deep Learning, Image Segmentation, Image Compression, U-Net, Autoencoder, CT Image, CNN.
A Sequential Clustering and Classification Framework for Lung Cancer Prediction
👥 Puneet Tiwari, Dr.Anupam Chouksey
📙 Abstract : Lung cancer is one of the major causes of cancer-related mortality worldwide, and early detection plays an important role in improving treatment outcomes. The increasing availability of medical imaging, clinical records, and patient-level datasets has created opportunities for applying machine learning and artificial intelligence to lung cancer prediction. Existing research has investigated supervised classification, deep learning, image segmentation, radiomics, and survival prediction; however, many approaches directly classify samples without first discovering natural patterns or subgroups in the data. This survey examines a sequential clustering and classification framework in which unsupervised clustering is first used to discover hidden structures within lung cancer data, followed by supervised classification for prediction. The survey discusses data preprocessing, feature extraction, clustering techniques, feature selection, classification algorithms, deep learning methods, performance evaluation, datasets, and challenges. Recent systematic reviews show substantial use of machine learning for lung cancer diagnosis and prognosis, while newer surveys emphasize the growing role of computational intelligence and deep learning in CT-based lung cancer analysis. The proposed sequential perspective can potentially improve subgroup discovery, reduce data complexity, and provide more interpretable prediction pipelines, although its effectiveness must be validated carefully against direct classification and end-to-end deep learning approaches.
🔖 Keywords :️ Lung Cancer, Machine Learning, Clustering, Classification, Deep Learning, CT Images, Feature Selection, Predictive Analytics, Artificial Intelligence, Medical Image Analysis.
BHUMIHARS: A Story of Lost Tribal Identity
👥 Shwetanshu Ranjan
📙 Abstract : This paper examines the sociological and constitutional basis for the claim that the Bhumihar community of Bihar and eastern Uttar Pradesh (Purvanchal) represents a case of "lost tribal identity" rather than an organically constituted upper caste. Using the five-fold criteria formulated by the Lokur Committee (1965) for the identification of Scheduled Tribes — primitive traits, distinctive culture, geographical isolation, shyness of contact, and general backwardness — the paper evaluates the historical and socio-economic profile of the Bhumihar community against each parameter. It argues that the community's association with forest-clearing and agrarian expansion in the hilly and plateau tracts of Gaya, Kaimur, Rohtas and Nawada during the colonial period, together with its documented exclusion from Brahmanical ritual status, endogamous isolation, and continuing economic backwardness, are consistent with a tribal origin obscured by colonial-era Sanskritization. Drawing a comparative parallel with the Meena community of Rajasthan — a Scheduled Tribe that nonetheless produced ruling lineages the paper contests the assumption that tribal status is incompatible with landholding, political power, or upper-caste social positioning. The paper situates this argument within the broader history of revisions to India's central and state Scheduled Tribe lists, and notes recent political articulations of the demand, including its invocation by Bhumihar youth on social media and by an elected legislator in the Bihar Legislative Assembly in 2026. The paper is offered as a contribution to academic debate rather than a definitive ethnographic claim.
🔖 Keywords :️ Bhumihar; Scheduled Tribe; Lokur Committee; Sanskritization; caste and tribe; Meena tribe; Bihar; Purvanchal.