Journal of Science Engineering Technology and Management Science

The development of artificial intelligence (AI) technology has had a significant impact on the architecture of the global economy. Emerging market economies are increasingly adopting AI technology to improve productivity, competitiveness, and innovation. However, the distributional impact of AI technology is yet to be investigated. This paper examines the impact of the adoption of AI technology on income inequality in emerging market economies from 2020 to 2024. This paper employs panel data analysis and fixed-effects models with robust standard errors to examine the impact of the adoption of AI technology on income inequality and whether higher education and financial system development can moderate this relationship. The results of this paper show that the adoption of AI technology results in an increase in income inequality in the short term. However, higher human capital and financial system development can moderate this relationship. The results of this paper show that technological development does not necessarily result in inclusive growth.

Keywords : Artificial Intelligence (AI), Income Inequality, Developing countries, Human Capital, Financial Development

Author : Ediga Ravindra Goud, Lavudya Bablu, Sagar Mekala

Title : Artificial Intelligence and Income Inequality in Developing Economies: The Moderating Role of Human Capital and Financial Development

Volume/Issue : 2026;03(05)

Page No : 1-8

This paper explores how social media marketing has interfered with routine consumer buying behavior in terms of products, purchase occasions as well as promotion schemes. Having employed a descriptive and causal-comparative design, the information was gathered through the structured questionnaire form, consisting of 426 online shoppers. Various analytical procedures such as exploratory factor analysis, Pearson correlation, multiple regression, and ANOVA were employed as study procedures and all the processes were run through IBM SPSS Statistics Version 26. The findings show that the graphic stimulating content and marketing plans play a great role in determining the level and nature of purchases. The strongest predictor of purchase frequency was content engagement (0.38, p < .01), whereas the pattern of usage on the platforms pointed at Instagram and YouTube as the most important channels of marketing. Respondents aged between 18 and 24 were more receptive to festival-based promotions, which was authenticated by ANOVA. The structural validity of the marketing dimensions was proved by factorial analysis. These data indicate that customized, topical, and visually engaging social media marketing would work to influence the everyday consumer purchase behaviour. The analysis has offered more empirical and strategic findings to the marketing efforts of marketers who would be interested in influencing the habitual buying behaviour through online means.

Keywords : Social media marketing, Consumer behavior, Promotional schemes, Routine purchases, Regression analysis

Author : Dr.S.Shalini

Title : IMPACT OF SOCIAL MEDIA MARKETING ON ROUTINE BUYING CONSUMER BEHAVIOUR

Volume/Issue : 2026;03(05)

Page No : 9-20

Malicious software known as Malware in the form of viruses, ransomware, and spyware has turned into a global epidemic, and research shows that the impact is intensifying. Numerous ways have been introduced to date to deal with these hazards. To handle this increasing problem, this paper proposes an effective Deep Neural Network (DNN) model that can be used to detect ransomware precisely. The model proves to be very effective in the separation of malicious and benign samples, with an accuracy, precision, recall, and F1-score of 99.76, and an AUC value of 0.98, which indicates the close to perfection of the classification. The findings reveal the high learning stability and generalization without overfitting, which is reinforced by the stable training and validation. Compared to the current methods, including KNN (83.9%), VGG-16 (90.5%), XGBoost (94.1%), and Logistic Regression (96%), the DNN-based model was better in its performance. On the whole, this paper highlights how deep learning can be used to reinforce cybersecurity protection and offer a scalable and intelligent method to counter ransomware attacks in the present digital environment.

Keywords : Ransomware Detection, Industrial Control Networks (ICNs), Machine Learning (ML), Cybersecurity, Anomaly Detection, Intelligent Systems, SCADA Security.

Author : Mr. Swapnil Joshi

Title : A ROBUST MACHINE LEARNING APPROACH FOR INTELLIGENT RANSOMWARE DETECTION IN INDUSTRIAL CONTROL NETWORKS

Volume/Issue : 2026;03(05)

Page No : 21-28

The electronic industry is the world's largest and most rapidly growing manufacturing industry. Because e-waste has a negative impact on human health, it must be repurposed in some way to be environmentally benign. Huge loads of electronic gadgets are sent across seas consistently, yet after their utilization, they make a muddled waste matter including an assortment of unsafe weighty metals, acids, toxic mixtures, and non-biodegradable polymers. E-waste is collected, dismantled, separated, and exported for recycling by a number of commercial companies. The compressive and tensile strengths of concrete specimens constructed from E-waste material will be tested in this study. The e-waste material is used to make coarse aggregates of 20mm in size. For a concrete mix of M30, multiple concrete mixes are made by replacing regular coarse aggregate with ewaste material in various percentages such as 0 percent, 20%, 40%, 60%, 80%, and 100%. Cement is replaced with Silica fume in percentages of 0% to 10% when coarse aggregate is replaced. The impact of replacing coarse aggregate with e-waste material and cement with Silica fume on concrete parameters such as compressive and tensile strength for hardened concrete at 28 and 90 days will be investigated. For each concrete mix, a certain number of examples are made.

Keywords :

Author : MULLA SAMAD1 , B FAYAZ2

Title : INFLUENCE OF E-WASTE AND SILICA FUME ON THE STRENGTH CHARACTERISTICS OF M30 CONCRETE

Volume/Issue : 2026;03(05)

Page No : 29-36

The purpose of this study is to investigate the effects of concrete on the environment there is a huge growth in Construction industry. Many new technologies have evolved very rapidly to cater different difficulties in the construction industry. Concrete is one of the most important materials in a construction industry, among all the material used for construction purposes. The process of manufacturing of cement emits near about eight to ten parent of total world’s carbon dioxide. The global warming gas is released when limestone and clays are crushed and heated to high temperatures. In recent year, the recycling of waste and industrial by products gaining popularity to make concrete environment friendly material and the concrete can be called as Green Concrete. Green concrete is defined as a concrete which uses waste material as at least one of its components, or its production process does not lead to environmental destruction, or it has high-performance and life cycle sustainability. In the present study an experimental investigation was conducted by using recycled coarse aggregates and industrial ceramic sludge material in M30 grade concrete. The percentage of ceramic sludge and RCA is 0%CS+0%RCA, 5%CS+5%RCA, 10%CS+10%RCA, 15%CS+15%RCA, 20%CS+20%RCA and 25%CS+25%RCA. The workability like sump and compaction factor is determined with various CS and RCA the strength values like compressive, split tensile and flexural strength is determined with various percentages of CS and RCA.

Keywords :

Author : Thalupuru Umesh1 G Rajesh2

Title : AN EXPERIMENTAL STUDY ON THE UTILIZATION OF INDUSTRIAL CERAMIC SLUDGE AND RECYCLED COARSE AGGREGATES FOR SUSTAINABLE GREEN CONCRETE

Volume/Issue : 2026;03(05)

Page No : 38-44

In this study, the fine and coarse aggregates were completely replaced by fly ash aggregates in fly ash concrete. A mix design was done for M40 grade of concrete by IS method. Ordinary Portland cement of 43 Grade was selected and fly ash aggregates were prepared by mixing fly ash with cement and water. The properties of fly ash fine aggregates and fly ash coarse aggregates were studied. The aggregate crushing value and aggregate impact value of fly ash coarse aggregates were also studied. The fly ash aggregates proportions of 0% , 10%, 20% , 30 % , 40% , 50% by aggregates weight were tried with a suitable water cement ratio 0.45 to get the fly ash aggregates. The concrete cubes, cylinders and beams were cast with the fly ash aggregates obtained from the above six cement fly ash proportions. Then the compressive strength, split tensile strength and flexural strength and durability were tested and compared with control concrete. This paper briefly presents the compressive strength development of fly ash aggregate concrete at different ages. The split tensile strength and flexural strength of all the concrete mix were also investigated at different days of curing.

Keywords : Fly ash aggregates (FAA), Compressive Strength, Split tensile strength, Flexural strength, etc.

Author : SHAIK MOHAMMAD WASEEM1 K BHAVANA2

Title : STUDY OF STRENGTH AND DURABILITY PARAMETERS OF CONCRETE MADE USING FLY ASH AGGREGATES

Volume/Issue : 2026;03(05)

Page No : 45-54

Peer-to-peer (P2P) lending is another important financial technology (FinTech) innovation, which has helped to connect the two parties (borrowers and lenders) using online platforms without an intermediary between them, a financial institution. This is because such platforms enable unsecured lending, as borrowers can take loans and investors can finance them based on their risk and return preferences. P2P lending has been experiencing a blistering evolution due to the digital transformation phenomenon, the need to seek alternative funding, and the adoption of new and effective technologies, including artificial intelligence (AI), machine learning (ML), and blockchain. Nevertheless, efficient credit risk analysis is one of the biggest issues associated with the sustainability of platforms and assists in decreasing the loan default rate. Older credit scoring models are often based on a very thin financial history and do not capture dynamic borrower behaviour, and are therefore not believed to be applicable to the new digital borrowing markets. FinTech solutions address such limitations with alternative data, predictive analytics and automated decision-making systems. The AI-based methods to optimize credit assessment, fraud detection, and dynamic risk profiling include random forests, gradient boosting, neural networks, and natural language processing. Another way these technologies drive financial inclusion is by empowering underserved individuals and small enterprises to access credit via digital lending ecosystems.

Keywords : Peer-to-Peer (P2P) Lending, FinTech, Credit Risk Assessment, Artificial Intelligence, Machine Learning, Financial Inclusion.

Author : Sandeep Shivam

Title : A REVIEW OF FINTECH INNOVATIONS IN PEER-TO-PEER LENDING AND CREDIT RISK ASSESSMENT TECHNIQUES

Volume/Issue : 2026;03(05)

Page No : 55-62

The phenomenon of cell imbalance usually affects the performance of multi-cell energy storage devices, where variations exist in terms of cell voltage and SoC. It causes low efficiency and poor utilization of battery energy. Therefore, this paper explores an active cell balancing scheme to improve energy distribution and utilization within a cell. This paper proposes and analyzes the simulation of a cell balancing algorithm via the use of the MATLAB/Simulink modeling software package, considering multiple cells with various initial SoCs. In order to equalize the charge distribution among different cells, the charging process involves transferring excessive charges from higher-energy cells to lower-energy cells. From the obtained results, it is clear that the disparity between the SoC values of cells reduces to zero, showing no sign of instability. Similarly, voltage differences have been minimized, which demonstrates successful redistribution. It is evident that the proposed algorithm performs efficiently and reliably since the differences between the SoC values of various cells have reduced significantly to the point of reaching equality

Keywords :

Author : P. Anil Kumar, V. Sree Charan, Qusai Yousuf Ahmed, M. Srinivas

Title : Design and Simulation of Active Cell Balancing using MATLAB/SIMULINK

Volume/Issue : 2026;03(05)

Page No : 63-74

In recent years, ensuring safety in public environments has become a critical challenge due to the rising number of weapon-related incidents. Traditional surveillance systems such as CCTV rely heavily on human monitoring, which is often inefficient due to fatigue, distraction, and delayed response. This project presents an intelligent deep learning-based system for dangerous object detection in public spaces to enhance security and reduce dependency on manual surveillance. The proposed system utilizes advanced object detection algorithms such as YOLO, Faster R-CNN, and SSD MobileNet to identify weapons like guns and knives in real-time from video streams. The system is trained on a diverse dataset consisting of weapon and non-weapon images, including confusion objects to reduce false positives. Image preprocessing techniques such as normalization, augmentation, and enhancement are applied to improve detection performance in challenging conditions like low lighting and crowded environments. The model evaluates performance using precision, recall, F1-score, and mean average precision metrics, ensuring accurate detection. Upon identifying a dangerous object, the system generates immediate alerts, allowing authorities to take timely action. This automated approach improves surveillance efficiency, reduces human workload, and enhances response time. The system is scalable, cost-effective, and suitable for deployment in areas such as airports, railway stations, malls, and educational institutions. Overall, the proposed solution contributes significantly to public safety by providing a reliable, real-time, and intelligent surveillance mechanism

Keywords :

Author : 1Mrs. B. NIRUPAMA, 2KONDRA ANJALI REDDY, 3KOPPELA PRANAVI, 4DONGALA MADHU

Title : DANGEROUS OBJECT DETECTION IN PUBLIC SPACES FOR SAFETY ASSURANCE

Volume/Issue : 2026;03(05)

Page No : 75-81

The rapid digital transformation of healthcare systems has highlighted the importance of secure, scalable, and interoperable solutions for managing patient data. Traditional paper-based medical records continue to dominate in many developing regions, leading to inefficiencies such as delayed diagnosis, fragmented patient history, and increased risk of data loss. Electronic Health Records (EHRs) offer a promising alternative; however, their adoption is hindered by challenges including limited infrastructure, high implementation costs, and critical concerns regarding data privacy and security. This project proposes a privacy-first cloud-based framework designed to enable secure sharing and management of EHR data across healthcare institutions. The system leverages cloud computing to provide centralized storage, real-time accessibility, and scalability while minimizing dependence on local infrastructure. Advanced security mechanisms such as encryption, secure authentication, and role-based access control are integrated to ensure confidentiality, integrity, and availability of patient data. Additionally, the framework supports interoperability, enabling seamless data exchange between healthcare providers. The proposed solution also incorporates audit trails and controlled access policies to enhance transparency and accountability. By addressing both technical and security challenges, the system aims to improve healthcare delivery, reduce redundancy, and enhance decision-making processes. Ultimately, this project contributes to building a reliable, efficient, and secure digital healthcare ecosystem suitable for resourceconstrained environments.

Keywords :

Author : 1Mr. R. BHARATH,2 CHIDURALA SWARAJ, 3SURUKUTLA MANOJ, 4BASA NARAHARI, 5KANDHATI VAMSHIDHAR REDDY

Title : HEALTH SAFE CLOUD - A PRIVACY FIRST CLOUD FRAMEWORK SECURE EHR SHARING

Volume/Issue : 2026;03(05)

Page No : 82-88

The Predictive Student Placement Recommendation System is an intelligent webbased application designed to assist students in evaluating their placement readiness and identifying suitable career paths using machine learning classification algorithms. In the modern competitive job environment, students often lack personalized guidance and data-driven insights to determine their strengths, weaknesses, and career direction. This system addresses these challenges by analyzing academic performance parameters such as SSC, HSC, degree percentage, MBA percentage, entrance test scores, and work experience to predict placement outcomes. In addition to placement prediction, the system evaluates skill-based attributes including programming ability, aptitude, problem-solving skills, project experience, abstract thinking, and design skills to recommend appropriate job roles such as Software Developer, Data Analyst, UI/UX Designer, Technical Support, and Technical Writer. The application integrates three major modules: Admin, Employer, and User, ensuring structured functionality and role-based access control. Machine learning models are trained on structured datasets and deployed within a Flask-based web application to provide real-time predictions and recommendations. Furthermore, the system integrates job portal functionality by displaying relevant job opportunities based on predicted roles, allowing students to apply directly. This integrated approach reduces uncertainty in career decisions, minimizes random job applications, enhances placement preparedness, and improves overall decision-making efficiency. The system provides a scalable, reliable, and user-friendly platform that bridges the gap between prediction systems and job portals, ultimately supporting both students and employers in achieving better placement outcomes.

Keywords :

Author : 1Mrs. L. SHIRISHA, 2PISKA PREETHI, 3KALWAKOLLU POOJA, 4SURUGU BHARATH RAJ, 5PASUNURI RISHI

Title : PREDICTIVE STUDENT PLACEMENT RECOMMENDATION SYSTEM USING MACHINE LEARNING CLASSIFICATION ALGORITHMS

Volume/Issue : 2026;03(05)

Page No : 89-96

Mental health disorders such as depression, anxiety, and stress have become significant global concerns due to rapid lifestyle changes, increased social pressure, and technological dependency. Early detection of such conditions remains challenging because many individuals hesitate to seek professional help or remain unaware of their mental health status. This project presents a machine learning-based system, MindPredictor, designed to classify mental health conditions using supervised learning algorithms by analyzing user-generated textual data, particularly from social media platforms. The system leverages natural language processing techniques, including text preprocessing, tokenization, stemming, lemmatization, and sentiment analysis, to extract meaningful insights from user input. Sentiment polarity and subjectivity scores are computed and used as features for classification. The processed data is then fed into machine learning models such as Naïve Bayes and hybrid classifiers to distinguish between depressive and non-depressive states. The system evaluates model performance using metrics like accuracy, precision, recall, and confusion matrix analysis to ensure reliable predictions. Results demonstrate that machine learning algorithms can effectively identify patterns related to mental health conditions and provide accurate classification outcomes. The proposed system offers a cost-effective, scalable, and accessible solution that can assist in early mental health assessment. It does not replace professional diagnosis but acts as a supportive tool for awareness and preliminary screening. By integrating artificial intelligence with healthcare, the system contributes to improved mental health monitoring, timely intervention, and reduced social stigma associated with psychological disorders.

Keywords :

Author : 1Mrs.L. SHIRISHA, 2MANISHAKUMARI, 3V. AISHWANTH, 4T. THARUN

Title : MIND PREDICTOR: MENTAL HEALTH STATUS CLASSIFICATIONUSINGSUPERVISEDMACHINE LEARNINGALGORITHMS

Volume/Issue : 2026;03(05)

Page No : 97-103

The Leather Vision system is an advanced artificial intelligence–based solution designed to automate the detection of surface defects in leather materials using deep learning techniques. Traditional leather inspection methods rely on manual observation, which is time-consuming, subjective, and prone to inconsistencies due to human fatigue and varying expertise. This project introduces a computer vision-driven approach that leverages convolutional neural networks (CNNs) to enhance accuracy and reliability in defect detection. High-resolution images of leather surfaces are captured under controlled lighting conditions and undergo preprocessing steps such as noise removal, normalization, and contrast enhancement to improve feature quality. A transfer learning-based model using MobileNetV2 is employed to extract deep features and classify defects such as cuts, scars, wrinkles, holes, and grain irregularities. The system is trained on a labeled dataset, enabling it to distinguish between defective and non-defective regions with high precision. Performance evaluation is conducted using metrics such as accuracy, precision, recall, and F1-score, demonstrating significant improvement over traditional inspection methods. The proposed system supports real-time defect detection, making it suitable for industrial applications where speed and consistency are critical. By reducing human dependency and minimizing material wastage, the system contributes to improved quality control and cost efficiency in leather manufacturing. Overall, this project highlights the transformative potential of artificial intelligence in modernizing industrial inspection systems

Keywords :

Author : 1Dr. S. SANJEEVA RAO, 2PASHAPU AKHILA, 3ANTHATI SHILPA, 4BADRI SANTHOSH

Title : LEATHER VISION: AI -BASED SURFACE DEFECT DETECTION IN LEATHER USING DEEP NEURAL INSPECTION SYSTEM

Volume/Issue : 2026;03(05)

Page No : 104-110

Road accidents are a critical public safety issue, especially in developing countries like India, where increasing vehicle density and poor traffic management contribute to high fatality rates. This project focuses on the development of an intelligent road accident analysis and accident severity prediction system using machine learning techniques. The system utilizes historical accident datasets to identify patterns and relationships among various factors such as road conditions, weather conditions, time of accident, driver behavior, vehicle type, and traffic density. Data preprocessing techniques including cleaning, normalization, and feature selection are applied to ensure accurate model performance. Multiple machine learning algorithms such as Decision Tree, K-Nearest Neighbors, Naïve Bayes, and AdaBoost are implemented and evaluated to determine the most effective model for predicting accident severity. The system classifies accidents into categories such as slight, serious, and fatal injuries, enabling better understanding of risk factors. Experimental results indicate that machine learning models outperform traditional statistical methods in predicting accident severity with higher accuracy. The proposed system provides data-driven insights that assist traffic authorities, policymakers, and emergency services in proactive decision-making and resource allocation. Furthermore, the system can be enhanced with real-time data integration and visualization dashboards to improve road safety strategies. Overall, this project contributes to reducing accident severity and improving transportation safety through intelligent prediction and analysis

Keywords :

Author : 1Mrs. K. KAVYA, 2P. HARSHITHA, 3Y. SRINIJA, 4R. AKHIL

Title : ROAD ACCIDENT ANALYSIS AND PREDICTION OF ACCIDENT SEVERITY

Volume/Issue : 2026;03(05)

Page No : 111-117

The rapid evolution of digital technologies has significantly influenced the transformation of traditional systems into more efficient and secure digital platforms. Voting, being a fundamental democratic process, still relies largely on conventional methods such as ballot papers and electronic voting machines, which suffer from issues like security vulnerabilities, voter impersonation, and limited accessibility. This project presents a secure virtual voting system that integrates facial authentication and blockchain technology to address these challenges. The system introduces a multi-level authentication mechanism that includes unique identification verification, voter credentials, and real-time facial recognition to ensure that only authorized individuals can cast votes. Facial recognition, powered by machine learning algorithms, enhances identity verification accuracy and prevents fraudulent activities. Blockchain technology is employed to store votes in a decentralized and tamper-proof ledger, ensuring transparency, immutability, and trust in the voting process. The proposed system enables remote voting, eliminating the need for physical presence at polling stations, thereby increasing voter participation and convenience. Additionally, the system provides real-time vote recording and result generation while maintaining data integrity and security. By combining biometric authentication with blockchain, the system significantly improves reliability, reduces human intervention, and minimizes operational costs. This approach modernizes the electoral process and offers a scalable, efficient, and secure solution suitable for future digital governance systems

Keywords :

Author : 1Mrs. NANDINI SREE, 2GATTIGORLA DEEKSHITHA, 3CHERLAKOLA SRIJA, 4MUPPARAPU SAI UDAY, 5MANUKONDA NAVEEN

Title : SECURE VIRTUAL VOTING SYSTEM USING FACIAL AUTHENTICATION AND BLOCKCHAIN

Volume/Issue : 2026;03(05)

Page No : 118-124

The rapid advancement of artificial intelligence in agriculture has enabled the development of intelligent systems for improving crop productivity and disease management. This project presents an AI-powered crop disease detection, forecasting, and management system that leverages deep learning and computer vision techniques to identify plant diseases from leaf images. Traditional disease detection methods rely on manual inspection, which is time-consuming, error-prone, and requires expert knowledge, limiting scalability in large agricultural environments. The proposed system overcomes these limitations by using pre-trained convolutional neural network models such as Xception and DenseNet121, fine-tuned for accurate classification of plant diseases. The system processes images through preprocessing steps including resizing, normalization, and feature extraction, followed by classification to determine disease type and severity. Additionally, it integrates environmental parameters such as temperature and humidity to forecast potential disease outbreaks, enabling proactive intervention. A web-based interface allows farmers to upload images and receive instant results along with treatment recommendations, improving accessibility and usability. The system enhances decision-making by providing timely alerts and reducing dependency on expert consultation. Furthermore, it promotes sustainable agriculture by minimizing excessive pesticide usage and improving crop yield. Overall, the proposed solution provides a scalable, efficient, and intelligent approach to crop health monitoring, contributing to food security and smart farming practices

Keywords :

Author : 1Ms. S. GUNASREE, 2M. NAMITHA, 3M. VIJAY KUMAR REDDY, 4T. JAYANTH

Title : AGRI HEALTH AI: AI-POWERED CROP DISEASE DETECTION, FORECASTING, AND MANAGEMENT SYSTEM

Volume/Issue : 2026;03(05)

Page No : 125-131

Cloud computing has become a dominant platform for storing and managing large volumes of data due to its scalability, cost-effectiveness, and accessibility. However, outsourcing sensitive data to third-party cloud service providers introduces significant security and privacy concerns. This project proposes an efficient and secure approach for cloud data storage, sharing, and retrieval using Attribute-Based Encryption (ABE). The system ensures data confidentiality by allowing only authorized users with matching attributes to access encrypted information. Unlike traditional encryption methods, which rely heavily on complex key management, the proposed system embeds access policies directly into the ciphertext, enabling fine-grained access control. Additionally, the integration of Attribute-Based Searchable Encryption (ABSE) allows users to perform secure searches over encrypted data without revealing sensitive information to the cloud provider. The system utilizes a hybrid encryption model combining symmetric encryption for data efficiency and ABE for secure key distribution. Furthermore, advanced cryptographic techniques such as bilinear pairings and Type-III curves enhance system security while maintaining computational efficiency. The proposed model improves scalability, reduces communication overhead, and supports dynamic user management. Experimental considerations demonstrate that the system is practical for real-world applications and ensures high security with minimal performance trade-offs. Overall, this approach provides a reliable solution for protecting sensitive cloud data while enabling efficient sharing and retrieval.

Keywords :

Author : 1Mrs. K. KAVYA, 2K. POOJITHA, 3B. INDHU, 4K. VINAY KUMAR

Title : ATTRIBUTE-BASED ENCRYPTION APPROACH FOR STORAGE, SHARING AND RETRIEVAL OF ENCRYPTED DATA IN THE CLOUD

Volume/Issue : 2026;03(05)

Page No : 132-138

The rapid growth of digital payment systems, ecommerce platforms, and online banking has significantly increased the volume of financial transactions, thereby elevating the risk of fraudulent activities. Traditional fraud detection systems, primarily based on static rule-based approaches, are no longer sufficient to handle evolving and sophisticated fraud patterns. This project presents a Cloud-Based Fraud Detection System designed to accurately identify fraudulent transactions using machine learning techniques. The system incorporates data preprocessing, feature engineering, and classification algorithms to analyze transaction patterns and distinguish between genuine and fraudulent activities. Various supervised learning models, including Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, Naive Bayes, and KNearest Neighbors, are implemented and evaluated using performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. To address the issue of imbalanced datasets, techniques such as resampling and Synthetic Minority Over-sampling Technique (SMOTE) are applied. The selected optimized model is integrated into a Flask-based web application that enables real-time fraud prediction. Cloud deployment ensures scalability, flexibility, and high availability, allowing the system to handle large volumes of transaction data efficiently. The proposed system enhances detection accuracy, reduces false positives, and provides a secure and reliable solution for online transaction monitoring. Overall, the integration of machine learning with cloud computing offers a robust framework for preventing financial fraud and improving trust in digital financial systems.

Keywords :

Author : 1Dr.M. RAMU, 2SIRI SATWIKA POLOJU,3 PUPPALASRINIJA, 4P KRANTHIKUMARREDDY

Title : CLOUD-BASEDFRAUDDETECTIONSYSTEM FORONLINETRANSACTIONS

Volume/Issue : 2026;03(05)

Page No : 139-144

Helmet Track AI is an advanced deep learningbased system designed to enhance road safety by automatically detecting helmet usage among motorcyclists in real-time traffic environments. The rapid increase in two-wheeler usage has significantly contributed to road accidents, particularly due to non-compliance with helmet regulations. To address this issue, the proposed system employs Convolutional Neural Networks (CNNs) integrated with Multi-Task Learning (MTL) to simultaneously perform motorcyclist detection and helmet classification. Unlike traditional systems that process detection and classification separately, this unified approach reduces computational redundancy and improves inference speed. The system is trained on diverse real-world datasets, enabling it to perform effectively under varying lighting conditions, weather scenarios, camera angles, and motion blur. It utilizes real-time video streams captured through surveillance cameras and processes them using advanced computer vision techniques for accurate detection. The integration of YOLO-based object detection further enhances performance by enabling fast and efficient localization of riders and helmets. Experimental results indicate that HelmetTrack AI achieves high precision, recall, and accuracy compared to conventional models. Additionally, the system supports edge deployment, making it suitable for smart city applications and automated traffic law enforcement. By providing real-time monitoring, violation detection, and data storage capabilities, the system contributes to improved compliance and reduced accident rates. Overall, HelmetTrack AI offers a scalable, efficient, and intelligent solution for modern traffic surveillance systems.

Keywords :

Author : 1Mrs. K. DURGA BHAVANI, 2K. VARSHA, 3G. RUTHVIK REDDY, 4A. SUVIKRAM

Title : HELMETTRACK AI: CNN-BASED MULTI TASKING HELMET DETECTION ON MOVING MOTORCYCLE

Volume/Issue : 2026;03(05)

Page No : 145-151

The rapid growth of mobile technology has transformed traditional service-based systems into efficient digital platforms. The Vehicle Rental Booking and Fleet Management System is designed as an Android-based application that simplifies and automates the vehicle rental process. The system provides a centralized platform where users can search, book, and manage rental vehicles, while vehicle owners can efficiently monitor and manage their fleet operations. The application is developed using Java for backend logic and XML for user interface design, ensuring a responsive and userfriendly experience. Firebase is used as a real-time database to store user details, vehicle information, booking records, payments, and reviews, enabling instant synchronization across all users. The system supports multiple roles such as Admin, Owner, User, and Driver, ensuring structured access control and improved system security. The Admin oversees system activities, vehicle owners manage bookings and vehicles, users can easily rent vehicles, and drivers handle ride operations. This system eliminates manual processes such as phone-based bookings and paper records, reducing errors and improving operational efficiency. Real-time updates prevent booking conflicts and ensure data accuracy. The application enhances transparency by allowing users to provide feedback and ratings. Overall, the system offers a scalable, secure, and efficient solution for modern vehicle rental services, improving user experience and operational management through digital transformation.

Keywords :

Author : 1Mrs. A. JYOSHNA, 2P. MANISH, 3S. CHARAN,4 V. VIVEKANANDA SWAMY

Title : RENT RIDE – VECHICLE RENTAL BOOKING AND FLEET MANAGEMENT SYSTEM

Volume/Issue : 2026;03(05)

Page No : 152-157

Money laundering is a critical financial crime that poses serious threats to economic stability and global financial systems. With the rapid growth of digital banking, online transactions, and financial technologies, detecting suspicious financial activities has become increasingly complex. Traditional Anti-Money Laundering (AML) systems primarily rely on rule-based approaches, which often fail to detect advanced laundering techniques and generate a high number of false positives. To address these limitations, this project proposes FinShield, an AI-driven system designed to detect money laundering activities within social transaction networks. The system integrates Machine Learning (ML), behavioural analysis, and network-based graph analysis to identify suspicious transaction patterns and hidden relationships among users. Instead of analysing transactions individually, FinShield evaluates the overall transaction ecosystem by modelling users as nodes and transactions as edges, enabling detection of circular transactions, layering patterns, and suspicious clusters. Multiple ML models such as Random Forest, Logistic Regression, and Support Vector Machines are used to classify transactions based on risk scores. The system also incorporates real-time monitoring and a web-based dashboard for visualization and alerts. By combining AI techniques with network intelligence, FinShield significantly improves detection accuracy, reduces false positives, and enhances decision-making for financial institutions. The proposed system provides a scalable, adaptive, and efficient solution for modern AML challenges, ensuring better financial security and fraud prevention in digital ecosystems.

Keywords :

Author : 1Mrs. L. SHIRISHA, 2D. SHIVANI, 3B. VIVEK, 4K. RAGHUNATH

Title : FINSHIELD : AI-DRIVEN DETECTION OF MONEY LAUNDERING SOCIAL NETWORK TRANSATIONS

Volume/Issue : 2026;03(05)

Page No : 158-163

The rapid advancement of digital technologies has transformed traditional processes into efficient and automated systems, particularly in the domain of elections. This project presents a secure and efficient blockchain-based online voting system designed to improve transparency, reliability, and accessibility in student elections. The proposed system eliminates the limitations of conventional paper-based voting, such as manual errors, time consumption, and lack of transparency. It enables students to cast votes remotely using authenticated login credentials, ensuring convenience and increased participation. Blockchain technology is integrated to provide a decentralized and tamperproof mechanism for storing votes, thereby enhancing data integrity and security. Each vote is recorded as a block in the chain, ensuring immutability and preventing unauthorized modifications. The system enforces a one-user-onevote policy and maintains voter anonymity while ensuring vote authenticity. Additionally, the platform includes an administrative module for managing candidates, scheduling elections, monitoring voting activities, and generating results automatically. The use of encryption, authentication mechanisms, and secure database management ensures confidentiality and protection against cyber threats. By automating the voting process, the system reduces human effort, eliminates paperwork, and delivers instant and accurate results. This project demonstrates how modern technologies like blockchain and web-based platforms can revolutionize election systems in educational institutions by making them more secure, efficient, and transparent.

Keywords :

Author : 1Mrs. V. LALITHA LAVANYA, 2K. DHANUSH KARTHIK,3 K. VISHAL GOUD, 4J. ABHISHEK REDDY, 5D. NIKITH RAJU

Title : VOTE – X: BLOCKCHAIN BASED ONLINE VOTING SYSTEM

Volume/Issue : 2026;03(05)

Page No : 164-170

Cloud computing has become a dominant paradigm for storing and managing large volumes of data due to its scalability, flexibility, and cost efficiency. However, outsourcing data to remote cloud servers results in the loss of direct control over data, raising critical concerns regarding data integrity, security, and trust. Traditional auditing mechanisms rely on centralized Third-Party Auditors (TPAs), which introduce risks such as single points of failure, lack of transparency, and potential collusion with cloud providers. To address these challenges, this project proposes a Blockchain-Based Decentralized Public Auditing (BDPA) system for secure cloud storage verification. The system leverages blockchain technology to eliminate dependency on centralized auditors by enabling decentralized verification through smart contracts. Users generate authentication tags using identity-based cryptography and upload data to the cloud. During auditing, a decentralized commit-reveal mechanism generates random challenges, and the cloud server responds with proof of data possession. The blockchain network verifies this proof without accessing actual data, ensuring privacy preservation. Audit results are permanently recorded on the blockchain, providing transparency and tamper-proof logging. The proposed system improves reliability, scalability, and resistance against malicious attacks while maintaining efficient verification with minimal communication overhead. Experimental evaluation demonstrates that the system effectively prevents data tampering, ensures secure auditing, and enhances trust in cloud storage services. Overall, integrating blockchain with cloud auditing provides a robust, decentralized, and privacy-preserving solution for modern data integrity verification

Keywords :

Author : 1Mr. JAHANGEER PASHA, 2G. VAISHNAVI, 3K. MANASA, 4G. ABHIRAM

Title : DATALOCK PRO: BLOCKCHAIN FOR SECURE CLOUD AUDITING

Volume/Issue : 2026;03(05)

Page No : 171-177

The rapid growth of Android applications has led to a significant increase in malware threats, posing serious risks to user privacy, data security, and device integrity. Traditional malware detection techniques often struggle to identify sophisticated and evolving threats due to their reliance on signature-based or static analysis methods. To address these challenges, this paper proposes a deep learning-driven pattern recognition framework integrated with an equilibrium optimizer for efficient Android malware detection. The proposed approach combines the feature extraction capabilities of deep learning models with the optimization strength of the equilibrium optimizer to enhance detection accuracy and model performance. The system begins by extracting relevant features from Android application packages (APKs), including permissions, API calls, and behavioral characteristics. These features are then processed using a deep learning model, such as a convolutional or recurrent neural network, to learn complex patterns associated with malicious and benign applications. The equilibrium optimizer is employed to fine-tune model parameters and optimize feature selection, reducing redundancy and improving classification efficiency. Experimental results demonstrate that the proposed hybrid model achieves high detection accuracy, precision, and recall compared to traditional machine learning approaches. The integration of optimization techniques significantly enhances model convergence and reduces computational complexity. Additionally, the system shows strong robustness against newly emerging malware variants, making it suitable for real-world deployment. Overall, this work provides an effective and scalable solution for Android malware detection by leveraging the synergy between deep learning and metaheuristic optimization, contributing to improved mobile security and threat mitigation.

Keywords : Android Malware Detection, Deep Learning, Equilibrium Optimizer, Pattern Recognition, Cybersecurity, Feature Optimization, Mobile Security, Machine Learning

Author : Siragoni Sai Divya, Dr A Yashwanth Reddy

Title : DEEP LEARNING-DRIVEN PATTERN RECOGNITION WITH EQUILIBRIUM OPTIMIZATION FOR ANDROID MALWARE DETECTION

Volume/Issue : 2026;03(05)

Page No : 178-184