International Journal on Recent and Innovation Trends in Computing and Communication
Not a member yet
8613 research outputs found
Sort by
AI Agents for Business Applications: A Review
Full text linkArtificial Intelligence (AI) agents have revolutionized business applications by automating processes, enhancing decision-making, and optimizing operational efficiency. This paper presents a comprehensive review of AI agents, categorizing their applications across domains such as customer relationship management (CRM), supply chain management, financial forecasting, and enterprise decision support systems. The evolution of AI agents from rule-based models to sophisticated multi-agent systems (MAS) and large language models (LLMs) has enabled businesses to leverage intelligent automation, real-time analytics, and predictive insights. AI-driven conversational agents have improved customer engagement, while AI-powered workflow automation has enhanced IT operations and security. Despite these advancements, challenges such as ethical considerations, security risks, interoperability, and long-term adaptability persist. This review synthesizes research contributions, identifying key strengths, limitations, and emerging research gaps in AI adoption for business. Future directions highlight the need for enhanced human-AI collaboration, standardization of AI agent interoperability, security-first AI architectures, and emotionally intelligent conversational systems. Addressing these challenges will ensure the responsible and effective deployment of AI agents, maximizing their transformative potential in business environments
"A Comparative Study of Behaviour Predictors for School Students in Indore Using Machine Learning Algorithms".
Full text linkPredicting student academic performance has become a key area in educational data mining, with machine learning techniques offering powerful tools for early intervention and decision-making. This study explores the application of classification models to forecast student success and behavioural outcomes, with the goal of improving academic support systems and reducing dropout rates. Two distinct datasets of student information were utilized, and three boosting-based machine learning algorithms - XGBoost, AdaBoost, and an Artificial Neural Network (DenseNet) - were implemented. Feature engineering techniques were applied to optimize input variables and enhance model effectiveness.
The results demonstrate that it is feasible to predict student behaviour and academic performance with significant accuracy using machine learning models. Among the evaluated methods, XGBoost and AdaBoost achieved the best predictive performance with an accuracy rate of approximately 88%. Conversely, the DenseNet-based neural network model produced the lowest accuracy, around 49%. These findings underscore the effectiveness of boosting methods for educational prediction tasks and highlight the role of machine learning as a practical approach to advancing educational research and institutional planning
Dynamic Multi-Cloud Security and Availability Optimization (DMCSAO) Algorithm for Overcoming Service Unavailability in Multi-Cloud Environments
Full text linkService unavailability in cloud computing environments poses significant challenges for organizations relying on cloud-based applications and services, leading to disrupted operations, financial losses, and compromised user experience. These challenges are particularly critical in sectors such as healthcare, finance, and e-commerce, where continuous service availability is essential for business operations and customer satisfaction. This research addresses these challenges with a novel Dynamic Multi-Cloud Security and Availability Optimization (DMCSAO) algorithm, designed to enhance service reliability and system resilience across multiple cloud providers. Comprehensive experimental analysis using network graph visualization and simulation techniques to evaluate system behavior under various failure scenarios, including network partitions and datacenter failures. The experimental framework tests different node densities (30, 50, and 100 nodes) and compares multi-cloud versus single-cloud deployments in real-world application scenarios. patterns and recovery strategies, while our network partition simulations show sub- linear response time scaling but exponential recovery time growth in larger deployments. The DMCSAO algorithm maintains high service availability during failure scenarios, compared to single-cloud results demonstrate substantial improvements in multi-cloud implementations, achieving reduction in average response times (23.5ms versus 27.8ms), lower packet loss rates (2.3% versus 3.8%), and fewer failover incidents. The visualization-based analysis reveals crucial insights into failure propagation environments. These findings provide practical guidelines for implementing resilient cloud security services and contribute significantly to the field of multi-cloud architecture optimization. Our research addresses critical challenges in cloud computing reliability and offers valuable insights for organizations adopting multi-cloud strategies, while also identifying important directions for future research in cloud security and availability optimization
Securing Digital Identities System through Blockchain Networks
Full text linkThe increasing importance of blockchain technology as an improvement in efficiency, security, and transparency across different fields has notably made a difference in identity and access management systems. Traditional security don’t extend sufficient protection, mainly because threats have become sophisticated. Decentralisation and cryptographic mechanisms assure the integrity of data and eliminate single points of failure offered by a blockchain. This paper highlights the transformative role of blockchain technology in cybersecurity, especially with focus on smart contracts, secure authentication techniques, and decentralised identity management. The paper provides an insight into technological advancements, challenges, and trends with respect to bolstering security and trust in online transactions. It goes a step further to evaluate the viability of implementing blockchain-based identity management systems by the corporate world and governmental organisations and takes into account factors including scalability, regulatory compliance, and user adoption.This paper also proposes blockchain based vehicle identity system using Practical Byzantine Fault Tolerance (PBFT) and Directed Acyclic Graph (DAG ) Consensu
Performance evaluation of ovarian cancer detection using on machine learning approaches based on feature selection
Full text linkOvarian cancer is one of the most dangerous genecology cancers because it does not show any symptoms in the early stages and there aren’t any good tests to find it. Early detection is important for increasing patient survival, but traditional diagnostic methods often don't have the right level of sensitivity and specificity. This study looks into how machine learning (ML) and deep learning (DL) can be used to find ovarian cancer early and accurately. We looked at five models: Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Network (ANN), and Elman Recurrent Neural Network (ERNN). We did this before and after using the RF algorithm to select features. The results show that feature selection made all of the models work much better. The ERNN model performed the best overall, with accuracy going from 89.8% to 92.5% and AUC-ROC going from 0.94 to 0.96 after feature selection. In the same way, ANN and RF got 92.1% and 91.0% accuracy, respectively, with big improvements in precision, recall, and F1-score. These results show how important it is to optimize features to make models work better. They also confirm that intelligent ML-based systems could be used to reliably find ovarian cancer early. 
Optimizing Cloud Data Management Through Oracle Database Cloud Engineering
Full text linkThe given paper will examine the possibilities of Oracle Database Cloud Engineering applying the strategies of performance, security, and compliance optimization of the cloud data management. It specifies on multi-layered security, such as encryption, access control, auditing, and monitoring based on AI in Oracle Cloud Infrastructure. Response time of queries, throughput and cost efficiency was analyzed by quantitative experimentation and simulation. Findings reveal that performance-optimized Oracle graphs contribute to maximizing the scale, minimizing latency, and high-BDPR and HIPAA adherence. It has been seen in the proposed framework that both security and efficiency can be maintained by using special attention to engineering optimization and automated deployment models with the integration of both -Oracle Cloud and on-premises systems
Smart Irrigation System using Raspberry Pi
Full text linkThis project focuses on automated and manual irrigation in addition with plant disease detection and growth monitoring using image processing on a Raspberry Pi 3B+. By leveraging TensorFlow Lite and OpenCV, the system can analyze plant health and trigger appropriate irrigation actions. The aim is to design accurate agriculture system by reducing water wastage and improving crop monitoring
Machine Learning-based Intrusion Detection System for Social Network Infrastructure
Full text linkThe growing number of cyber-attacks demands a critical measure to prevent unauthorized data access. Thus, intrusion detection has become critical to deal with such attacks. This work attempts to identify malicious connections using a few key parameters. The system has been trained using data relating to normal and abnormal events through machine learning and data mining techniques. To detect intrusions, this study assessed five distinct machine learning models: Random Forest, Bagging, Boosting, Support Vector Machine, and K-Nearest Neighbor (KNN). Based on the number of features, iterations, and hyperparameters, the models were evaluated using experimental data collected in real time. With a detection rate of up to 98.7%, the Random Forest approach surpassed existing machine learning models for intrusion detection. The paper proposes a novel intrusion detection system (IDS) based on these findings that successfully identifies possible threats before they seriously compromise network security and stop cyberattacks
Self-Supervised Hierarchical Representation Learning for Multi-Dimension Context
Full text linkSelf-supervised hierarchical representation learning offers an effective approach to capturing multi-dimensional context from unlabeled data. A key challenge in representation learning is integrating information from diverse aspects of the input, particularly when labeled data is limited. To address this, a novel strategy can be introduced that learns representations hierarchically, enabling the capture of context at varying levels of abstraction and across multiple dimensions. The process begins by modeling different contextual facets through component-specific representations, each capturing distinct semantic and structural attributes. A dynamic aggregation mechanism then combines these representations in a hierarchical manner, allowing information to propagate across levels of contextual abstraction. This enables the encoding of both fine-grained nuances and broader contextual dependencies. By leveraging self-supervised learning, the approach optimizes for inherent relationships within the multi-dimensional context, enabling the acquisition of robust representations from unlabeled data. This makes it particularly suitable for domains where labeled data is scarce or costly to obtain. Experimental results highlight the ability to learn rich, hierarchical representations that enhance performance on downstream tasks requiring deep contextual understanding. Key technical contributions include: (1) a context-aware masking strategy using Text Encoder for semantic recovery of masked fields, (2) a Hierarchical Model that fuses fine-grained tabular features with coarse-grained concepts and (3) a multi-stage training code base combining contrastive loss for cross-document alignment (RFP-bid pairs) and silhouette scores (from scikit-learn) to validate cluster coherence
Literature survey on Feature Extraction methods using CBIR Visual Search
Full text linkEfficient image retrieval relies on robust feature extraction methods capable of capturing the distinct characteristics of color, texture, and shape. This study investigates diverse techniques across these domains, emphasizing their impact on the ranking accuracy of retrieved images. In the color domain, methods such as Color Moments (CM), Color Moment Invariant Model (CMI), Dominant Color-Based Vector Quantization (DCVQ), MPEG-7 Dominant Color Descriptor, and integrated color-texture approaches are explored for their precision in identifying chromatic variations. Texture feature extraction techniques, including Discrete Wavelet Transform (DWT), Statistical Edge Detection (SED), Modified Scalable Descriptor (MSD), and Local Derivative Radial Patterns (LDRP), alongside Support Vector Machine (SVM) classifiers, are assessed for their ability to identify and rank images based on structural complexity. For shape features, advanced techniques such as boundary moments, complex coordinates, curvature scale space, intersection point mapping, and merging strategies are evaluated for their role in preserving spatial and geometric fidelity. By examining these methods in the context of top-ranked image retrieval, this work provides a comparative framework to guide the selection of optimal feature extraction techniques for high-performance image analysis systems