A Study on the Role of Small Finance Banks in Promoting Financial Inclusion
Authors: Kummari Mahesh, Mr.Bh.L.Mohan Raju
Abstract: Financial inclusion has become one of the most important drivers of inclusive economic growth, particularly in developing economies like India. A large segment of the Indian population, especially those living in rural and semi-urban areas, has traditionally been excluded from formal financial systems due to various barriers such as low income, lack of financial awareness, inadequate banking infrastructure, and reliance on informal sources of credit. These challenges have led to financial inequality and limited economic opportunities for underserved populations. In order to address these issues and promote inclusive growth, the concept of financial inclusion has gained prominence, focusing on providing affordable and accessible financial services to all sections of society. In this context, Small Finance Banks (SFBs) were introduced as a significant initiative to enhance financial inclusion. These banks are specifically designed to serve the needs of low-income groups, small business owners, micro-entrepreneurs, and rural populations. The primary objective of SFBs is to provide basic banking services such as savings accounts, deposits, loans, and digital banking facilities to individuals who are often neglected by traditional commercial banks. By targeting underserved segments, SFBs play a crucial role in bridging the gap between formal financial institutions and economically weaker sections of society.
A Study On Consumer Perception Towards Food Delivery Apps With Special Reference To Swiggy
Authors: Sriyadoju Bhanu Prasad, Dr. Vellala Subrahmanya Ramamurty
Abstract: This study looks at how consumers feel about online meal delivery services, specifically focusing on Swiggy. Consumer purchasing behavior has changed dramatically as a result of the quick development of digital technology and the rising use of smartphones, especially in the food service sector. Because they provide convenience, variety, time savings, and simple payment options, online meal delivery services like Swiggy have become an essential part of urban lifestyles. This study's main goal is to examine the variables affecting Swiggy users' preferences and satisfaction. Convenience, cost, delivery time, service quality, special offers, user interface, and customer service are among the factors that are the focus of the study. Additionally, it seeks to comprehend customer perception, usage frequency, and general satisfaction levels. The study's conclusions aid in comprehending customer behavior and expectations regarding online meal delivery services. Insights from the study can help businesses like Swiggy maintain a competitive edge in the expanding online meal delivery market by raising customer satisfaction, improving service quality, and creating efficient marketing plans.
DOI: http://doi.org/10.5281/zenodo.21352802
Investment Behavior of Investors
Authors: Mudarapu Pavan, Mr.Bh.L.Mohan Raju
Abstract: This study examines the study on Investment Behavior of Investors at Union Bank of India at Hyderabad, focusing on how financial strategies are designed to optimize returns while minimizing risks. Portfolio management plays a crucial role in guiding investors toward effective allocation of assets across various financial instruments such as equities, bonds, mutual funds, and other securities. The research aims to analyze the methods adopted by Union Bank in constructing and managing investment portfolios, considering factors like risk tolerance, market trends, diversification, and return expectations. The study adopts both qualitative and quantitative approaches, using primary data collected through surveys and secondary data from financial reports and publications. It evaluates key portfolio management techniques such as diversification, asset allocation, risk assessment, and performance evaluation. The findings indicate that structured portfolio strategies significantly influence investment decisions and enhance financial outcomes for clients. Furthermore, investor awareness, financial goals, and market volatility are identified as critical determinants in decision-making.
DOI: http://doi.org/10.5281/zenodo.21353589
Gold Exchange Traded Funds in Net Worth Stock Broking Limited
Authors: Bachalakuri Naveen, Dr. K. Pushpa Latha
Abstract: An exchange traded fund (“ETF”) is a form of securities that tracks an index, sector, commodity, or other asset and may be bought and sold on a stock exchange much like a regular stock. An ETF can be set up to track anything from a single commodity's price to a huge and diverse group of securities. ETFs can even be built to follow certain investment strategies. The need and scope of the study is India is the world's largest gold-consuming country. China, on the other side, has the world's fastest-growing economy. The need and scope of the Both India and China are in the process of liberalizing restrictions governing the import and selling of gold in order to facilitate massive gold purchases. In India, online commodity trading is relatively new compared to the stock market. Online commodities trading is dominated by four exchanges: Multi Commodity Exchange of India Limited (“MCX”), National Commodity & Derivatives Exchange Limited (“NCDEX”), National Board of Trade (“NBOT”), and National Multi-Commodity Exchange of India Limited (“NMCE”). As a result, the study scope of the commodities market is relatively broad in the market, which is mostly focused on gold An analysis is carried out using performance evaluation techniques such as Standard Deviation (“STDV”), Variance, Covariance, Correlation and BETA value. Data for this study have been collected from 2017 to 2021. This research enables market analysts and investors who find the best outlook in the Gold ETF’s.
DOI: http://doi.org/10.5281/zenodo.21354433
A Study on Customer Behaviour Towards Bajaj Electronics
Authors: Thumma Navin, Associate Professor Dr. Vellala Subramanya Rama Murty
Abstract: This work looks into how people act when shopping at Bajaj Electronics, a top name in Indian consumer tech stores. With an aim to grasp what buyers like, how they shop, how happy they are, along with what sways their choices there. Examining things such as how well items perform, cost structure, ads and discounts, feel of the outlet, support given by staff, if products are in stock, plus public view of the brand. Each piece adds context to why shoppers respond the way they do. From customer surveys at Bajaj Electronics came the core details driving this work. Information beyond firsthand replies arrived via published material – books, articles, internal documents, online sources – all tied to buying habits in stores. To make sense of numbers, methods like share calculations and visual charts took center stage. Patterns emerged only after sorting results using these techniques. Most people choose Bajaj Electronics because it carries many items, sets fair prices, runs eye-catching deals, while delivering steady help after buying. Product performance matters just as much as how well-known the name is, along with savings offered, and whether staff respond when needed. Online promotion stands out too, especially through platforms where users connect daily, shaping who stops by and what they notice.
A Study on Referral Marketing by Doctors and its Impact on Lab Revenue with Reference to Tesla Unipath (Pvt.Ltd)
Authors: Kalval Kumar Swami Chary, Associate Professor Dr. Sivaji Jinka
Abstract: Referral marketing by doctors plays a significant role in the growth and profitability of diagnostic laboratories. Doctors often recommend specific diagnostic centers to patients based on factors such as service quality, accuracy of test results, reliability, turnaround time, and professional relationships. These referrals directly influence the number of patients visiting a laboratory and, consequently, its revenue. Understanding the impact of doctor referrals helps diagnostic laboratories develop effective marketing strategies while maintaining ethical healthcare practices. This study focuses on the impact of doctor referral marketing on the revenue of Tesla Unipath (Pvt.) Ltd. The primary objective is to examine how referrals from doctors contribute to patient inflow and revenue generation. It also evaluates the factors that influence doctors' referral decisions, the level of satisfaction with the laboratory's services, and the challenges faced in maintaining referral relationships. The study is based on both primary and secondary data. Primary data were collected through structured questionnaires and interviews with doctors and laboratory personnel, while secondary data were obtained from company records, journals, books, and relevant websites. Descriptive statistical tools such as percentages, tables, and charts were used to analyze the collected data. The findings indicate that doctor referrals have a substantial positive impact on the laboratory's revenue by increasing patient volume and strengthening the laboratory's market reputation. The study also reveals that accurate test reports, timely report delivery, advanced diagnostic technology, quality customer service, and strong professional relationships are the key factors influencing doctors' referral decisions. The study concludes that maintaining ethical referral practices, continuously improving service quality, adopting modern diagnostic technologies, and strengthening communication with healthcare professionals can enhance patient trust and contribute to sustainable revenue growth for Tesla Unipath (Pvt.) Ltd.
Mutual Fund Portfolio and Risk Assessment with Reference to HDFC Bank: An Analytical Study
Authors: Uppunuti Sai Charan, Associate Professor Dr. M.P. Suri Ganesh
Abstract: Mutual funds have emerged as one of the most preferred investment avenues for individual and institutional investors, offering diversification benefits, professional management, and relatively lower risk compared to direct equity investments. This study analyzes the mutual fund portfolio and associated risks with special reference to HDFC Bank — one of India's leading private sector banks and a major distributor of mutual fund products. Using secondary data spanning FY 2021–22 to FY 2025–26, the study evaluates four representative HDFC mutual fund schemes — Large Cap, Mid Cap, Focused, and Dynamic Debt — through key financial metrics including average annual return, Compound Annual Growth Rate (CAGR), standard deviation, beta, Sharpe ratio, and coefficient of variation. The findings reveal that equity-oriented schemes, particularly HDFC Mid Cap Fund and HDFC Focused Fund, significantly outperformed the benchmark index over the study period, though with higher volatility. HDFC Dynamic Debt Fund demonstrated stability and defensive characteristics suitable for conservative investors. The study concludes that a diversified portfolio approach combining growth and stability-oriented schemes offers the most balanced risk-return outcome for investors.
Deep Learning – Based Satellite Image Segmentation
Authors: Assistant Professor B.Naresh, D. Swasthitha
Abstract: Semantic segmentation of satellite imagery has become a fundamental task in remote sensing because it enables accurate identification of land-use patterns for applications such as urban planning, environmental monitoring, disaster management, and resource assessment. However, the complex spatial characteristics of satellite images, including varying resolutions, atmospheric interference, and class imbalance, make precise pixel-level classification a challenging problem. This study presents a modified U-Net-based deep learning framework for multi-class semantic segmentation of satellite images using the Dubai Satellite Imagery Dataset. The proposed model classifies six distinct land-use categories, namely buildings, roads, vegetation, water bodies, barren land, and miscellaneous regions. The preprocessing pipeline incorporates image patch generation, MinMax normalization, one-hot encoding, and extensive data augmentation to improve training efficiency and enhance model generalization. To address the imbalance among land-cover classes, the framework combines Dice Loss with Categorical Focal Loss, enabling improved segmentation performance for both majority and minority classes. Model optimization is performed using the Adam optimizer, while segmentation quality is evaluated through Accuracy, Intersection over Union (IoU), and Jaccard Index. Experimental results demonstrate that the proposed approach achieves reliable segmentation performance across diverse satellite scenes, making it suitable for practical remote sensing applications. The study further discusses strategies for handling spatial resolution challenges and improving segmentation robustness in complex urban environments.
An Intelligent Multi-Agent AI Framework for Automated Candidate Interview Assessment
Authors: Assistant Professor B.Saritha, B.Sai keerthana
Abstract: Organisational development relies on recruiting and applicant evaluation. However, traditional interview techniques include manual coordination, static questionnaires, and subjective evaluations, which prolong the recruitment process and lead to inconsistent candidate selection. For screening candidates, arranging interviews, evaluating responses, and reporting performance, current recruiting systems often rely heavily on human engagement, which makes them inefficient when dealing with large amounts of applications. In addition, recruiters and applicants are both severely impacted by a lack of transparency caused by delayed response and restricted automation. In order to circumvent these restrictions, this research exposes an AI-driven online interview platform that automates the whole interview process using smart software agents and cutting-edge AI methods. With the suggested platform, a unified recruiting environment may be achieved via applicant registration, administrator permission, question development using AI, automatic response assessment, performance monitoring, and score alerts sent via email. The system uses a combination of Artificial Intelligence (AI), Bidirectional Long Short-Term Memory (BiLSTM) models, Convolutional Neural Networks (CNN), Speech-to-Text (STT), Text-to-Speech (TTS), and Natural Language Processing (NLP) to analyse candidate responses during dynamic interviews and assess sentimental and emotional traits. Intelligent interview management and objective candidate assessment are accomplished through the collaborative efforts of a multi-agent architecture that includes the Question Management Agent (QMA), Response Management Agent (RMA), Multimodal Emotion and Sentiment Analyser Agent (MESAA), and Comprehensive Evaluation and Scoring Agent (CESA). Emotion identification, sentiment analysis, and question management are all handled by the framework using datasets such as SAVEE, TESS, CREMA-D, ISEAR, and MedMCQA. The metrics BLEU, ROUGE, BERTScore, and Completeness Score are used to assess the candidates' answers. In contrast to the CNN model's 95% training accuracy and 67% validation accuracy for audio emotion identification, the BiLSTM model only manages 70% training accuracy and 60% validation accuracy when it comes to text emotion recognition, according to experimental assessment. The suggested platform offers a scalable, efficient, and impartial answer to contemporary recruiting by combining intelligent automation with thorough applicant assessment; all while maintaining the original methodology, architecture, datasets, and trial results.
Machine Learning-Based Multi-Factor Authentication for Secure Cardless Financial Services
Authors: Assistant Professor G.Chandram, Bhargavi Sammeta
Abstract: The rapid growth of digital banking and electronic payment services has increased the demand for secure and reliable authentication mechanisms capable of protecting financial transactions from fraud and unauthorized access. Conventional card-based banking systems remain vulnerable to card theft, skimming attacks, credential compromise, and identity fraud. To address these security challenges, this paper presents an intelligent cardless financial transaction framework that combines Machine Learning (ML) with Multi-Factor Authentication (MFA) to provide a secure and efficient user authentication process. The proposed system utilizes facial recognition as the primary biometric verification method, where facial features are extracted using a deep Convolutional Neural Network (CNN) based on the pretrained ResNet-50 model. Sensitive biometric information and One-Time Passwords (OTPs) are protected using the Advanced Encryption Standard (AES) algorithm to ensure secure storage and transmission. After successful facial verification, an OTP is generated and verified as an additional authentication factor before authorizing the financial transaction. This multi-layered authentication strategy significantly enhances protection against identity theft, fraudulent transactions, and unauthorized account access while maintaining a user-friendly banking experience. The integration of deep learning, biometric authentication, encryption, and OTP verification provides a scalable and robust security framework for next-generation digital banking systems. The proposed approach demonstrates the potential of intelligent authentication technologies to improve transaction security, operational efficiency, and customer trust in modern financial services.
A Logistic Regression-Based Predictive Model for Customer Purchase Intent Analysis
Authors: E.Poorna, B. Vaishnavi
Abstract: Predicting whether or not a consumer will make a purchase is a crucial part of contemporary business intelligence since it allows companies to tailor their marketing efforts to each individual customer. When faced with massive amounts of changing consumer data, traditional rule-based techniques often fail in their capacity to adapt and make accurate predictions. Using demographic, transactional, and behavioural characteristics, this research proposes a Logistic Regression-based machine learning framework for predicting consumer purchase intent. Age, browsing history, prior purchases, product interactions, and engagement metrics are some of the customer-related variables that the suggested system uses to assess the likelihood of a purchase. The prediction pipeline incorporates thorough preprocessing approaches to enhance data quality and model efficacy. These techniques include managing missing values, One-Hot Encoding, feature normalisation, and feature selection. Optimising model performance while preserving computational efficiency and interpretability is achieved by hyperparameter adjustment. Standard performance criteria such as accuracy, precision, recall, and F1-score are used to assess the usefulness of the suggested technique. The results show that it can give solid purchase predictions. Businesses may find new clients, tailor ads to each individual, and increase interaction with existing consumers all with the help of the built framework's real-time prediction capabilities. The results of the experiment show that Logistic Regression is a great tool for predicting if a consumer will make a purchase, which may help with data-driven decision-making and boost growth in the long run.
An Intelligent Image-Based Two-Factor Authentication System for Enhanced User Security
Authors: Assistant Professor G.Chandram, Bhargavi Sammeta
Abstract: The necessity for trustworthy and secure user authentication methods has grown dramatically due to the quick expansion of online services. Traditional password-based authentication techniques are susceptible to security risks such phishing, brute-force attacks, credential theft, and password reuse. This paper proposes a safe two-factor authentication (2FA) architecture that combines customized photo verification with conventional login credentials to solve these issues. Users provide basic login credentials and choose one or more personal photos, each linked to a distinct keyword, during user registration. One of the registered keywords is shown at random by the system once the username and password have been verified during the login procedure. The matching customized image must then be uploaded by the user as the second authentication factor. The Perceptual Hashing (pHash) technique, which effectively evaluates image similarity while retaining robustness against small image alterations, is used in image verification. The framework enhances system security by reducing the possibility of unauthorized access through password-guessing attacks by applying distinct authentication attempt limitations for recognized and unrecognized devices. Without the need for extra hardware or intricate verification processes, the suggested authentication paradigm improves account security. The framework is appropriate for contemporary web applications, cloud platforms, financial systems, educational portals, and enterprise authentication environments because it strikes a balance between security, usability, and computational efficiency by fusing personalized visual authentication with perceptual image hashing.
A Blockchain-Enabled Secure Online Voting Framework with Aadhaar Authentication and Real-Time Facial Verification
Authors: Assistant Professor M.Sai Manasa, B. Swapna
Abstract: With the fast evolution of digital technology, it is more critical than ever to ensure that election procedures are secure, transparent, and accessible. Traditional voting procedures, which include paper ballots and Electronic Voting Machines (EVMs), have high operating expenses, need a large number of workers, cause votes to be counted later than expected, and are difficult for some voters, such Non-Resident Indians (NRIs), to access. Presented here is a secure blockchain-enabled online voting framework that, in response to these issues, combines cutting-edge web technologies with sophisticated authentication procedures to provide a trustworthy and impenetrable voting space. To guarantee voter authentication, session integrity, and safe vote recording, the proposed system integrates Aadhaar-based OTP verification, MTCNN-based facial recognition, MobileNetV2-powered real-time face monitoring, and smart contracts based on the blockchain. In order for the system to identify suspicious activity and immediately reject compromised voting sessions, it is constantly monitored. This monitoring includes live session validation and screen surveillance. Using end-to-end encryption enhances the security of votes and keeps data intact all the way through the election. The results of the experiments show that the suggested digital voting system improves upon traditional voting methods in terms of accessibility, security, and transparency. An estimated 5% increase with the introduction of online voting is anticipated as a result of the implementation, bringing the participation of NRI voters to about 72%. The assessed statistical significance is 0.049. A digital voting infrastructure that is scalable, safe, and efficient may be built using Blockchain, MobileNetV2, and TensorFlow. This infrastructure will be able to support future democratic processes and encourage inclusive electoral participation.
Hybrid Speech-Based Gender Recognition and Age Classification Using LSTM Networks
Authors: S.Akhila, Ch.sathvika
Abstract: Due to its extensive usage in intelligent virtual assistants, healthcare systems, customer analytics, and human-computer interface, voice-based demographic prediction has become a significant field of study in speech processing. Developing personalised and context-aware services is made possible by accurately recognising demographic factors like age and gender from voice signals. For the purpose of determining an individual's age and gender from audio recordings, this research introduces a framework that combines deep learning with machine learning. To assess the benefits of sequential deep learning over traditional machine learning methods, we compare Logistic Regression's performance with that of an LSTM network for gender categorisation. Logistic Regression is used for multi-class classification after K-Means clustering is used to create synthetic age groups for age prediction. To accurately describe speaker-specific speech patterns, the suggested framework employs a number of spectral and acoustic parameters, such as statistical frequency descriptors, dominant frequency, interquartile range, spectral centroid, and mean fundamental frequency. In order to make the model more resilient and the predictions more accurate, we use thorough preprocessing, feature scaling, and feature selection. In terms of gender classification, the experimental evaluation shows that the LSTM model outperforms the baseline Logistic Regression model with a 98.5% accuracy rate, while the hybrid age prediction method successfully categorises synthetic age groups with high predictive performance. These results show that for intelligent applications in the real world, combining deep learning with traditional machine learning gives a scalable and dependable solution for voice-based demographic prediction.
An Ensemble Machine Learning Framework for Accurate House Rent Prediction
Authors: Assistant Professor S.Venkateswara Rao, C. Nikitha
Abstract: The ever-changing real estate market and the ever-increasing density of metropolitan areas have made accurate home rent forecast a must-have tool for investors, renters, property owners, and real estate agents. Historical patterns, manual estimate, and expert opinion are the mainstays of conventional pricing approaches, yet they often overlook the intricate interrelationships between various property qualities. This research suggests a combination of the Random Forest and Extreme Gradient Boosting (XGBoost) regression models to overcome these shortcomings and provide reliable rental home predictions. With the help of thorough data preparation, feature engineering, normalisation, and categorical encoding, the suggested framework enhances prediction performance using the Housing Price in India dataset, which contains 193,011 property records. The prediction power of the regression models is increased by adding additional parameters like bedroom-to-bathroom ratio and price per square foot. Highly accurate rental price predictions are generated by combining the strengths of the Random Forest and XGBoost models, which are trained on the processed dataset. In addition to Accuracy, Precision, Recall, and F1-score, the following metrics are used to assess the model's performance: Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared (R²). Overall, the experimental findings show an impressive performance in terms of prediction accuracy (around 98.7%), precision (0.987), recall (0.985), and F1-score (0.986), with an MSE of 0.0001, RMSE of 0.01, MAE of 0.01, and R³ of 0.9998. While maintaining the integrity of the dataset, algorithms, methodology, and experimental results, the suggested ensemble learning framework offers a dependable, efficient, and scalable answer to the problem of intelligent home rent prediction.
An Intelligent Machine Learning Framework for Accurate Ocean Wave Energy Forecasting Using Deep Learning and Hybrid Prediction Models
Authors: Assistant Professor Seemala Gouthami, P.Rama Krishna
Abstract: The growing demand for clean and sustainable energy has increased the importance of ocean wave energy as a reliable renewable power source. Efficient utilization of this resource depends on accurate forecasting of wave characteristics, which directly influences energy conversion efficiency, operational planning, and power grid stability. Conventional numerical forecasting models often struggle to represent the highly dynamic and nonlinear behavior of ocean waves, resulting in reduced prediction accuracy under changing environmental conditions. This paper presents a comprehensive study of modern Machine Learning (ML) techniques for wave energy forecasting, emphasizing the application of Deep Learning (DL) architectures, ensemble learning methods, and hybrid prediction models. Various learning approaches, including Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), Random Forest (RF), XGBoost, and transformer-based models, are analyzed for their ability to estimate significant wave height (SWH), wave period, and wave energy potential. The study also examines hybrid frameworks that integrate physical oceanographic models with data-driven learning techniques to improve forecasting reliability under complex marine conditions. Comparative evaluation demonstrates that hybrid deep learning models generally achieve better prediction accuracy, lower forecasting error, and greater robustness than conventional machine learning approaches. Furthermore, recent developments in transformer networks, Generative Adversarial Networks (GANs), and real-time data assimilation are discussed as promising directions for improving future wave energy prediction systems. The findings indicate that intelligent forecasting frameworks can significantly enhance renewable energy management, optimize wave energy converter performance, and support efficient integration of ocean energy into modern power systems.
Real-Time Virtual Mouse System Using Vision-Based Hand Gesture Recognition
Authors: Assistant Professor Shaik Shireenbanu, Y.Vasudha
Abstract: The increasing usage of computer vision technologies, which allow for touchless interaction between humans and computer systems, has drastically altered human–computer interaction (HCI). Because they require additional equipment and physical engagement, conventional mouse devices may be less useful in settings where hands-free operation is preferred. This work proposes a vision-based virtual mouse system that uses real-time hand gestures recorded by a conventional camera to enable users to perform standard mouse tasks. The suggested system uses computer vision techniques to watch and identify hand movements and associates specific gestures with scrolling, double-clicking, left-clicking, and right-clicking. Users can interact with the operating system without a physical mouse by identifying motion patterns in video frames captured by the camera. The technology offers dependable gesture recognition and fast cursor control, and experimental testing shows that it works well with straightforward backgrounds and appropriate lighting. Presentations, smart classrooms, medical settings, assistive technology, and touchless computing systems could all benefit from the suggested approach, which provides a low-cost, device-free substitute for desktop interaction. Additionally, the study outlines future improvements to increase robustness and real-time performance while discussing the limitations of vision-based interaction in difficult environmental scenarios.
Generative AI-Driven Threat Intelligence for Intelligent Cybersecurity Defense
Authors: Assistant Professor Singaram Navya, S.Venkateswara Rao
Abstract: The increasing frequency and sophistication of cyberattacks have created significant challenges for organizations seeking to protect their digital infrastructure and sensitive information. Conventional cybersecurity solutions often rely on predefined rules and signature-based detection methods, which may struggle to identify emerging and previously unknown threats. To overcome these limitations, this paper presents an intelligent threat intelligence framework that leverages Generative Artificial Intelligence (Generative AI) to strengthen cyber defense mechanisms and improve proactive threat detection. The proposed framework integrates deep learning models, generative techniques, and automated threat analysis to continuously monitor security events, identify abnormal behavioral patterns, and generate actionable threat intelligence in real time. By analyzing large volumes of structured and unstructured security data, the system can recognize evolving attack strategies, predict potential vulnerabilities, and support rapid decision-making for security analysts. Furthermore, the framework facilitates automated incident analysis, threat classification, and adaptive response generation, thereby reducing detection latency and improving the overall efficiency of cybersecurity operations. The performance of the proposed model is evaluated by comparing it with existing approaches, including DCGAN, WGAN, BERT, and SPADE, using performance metrics such as accuracy, false positive rate, false negative rate, and response time. Experimental observations indicate that the proposed Generative AI framework provides superior threat detection capability, faster response, and improved reliability compared with conventional methods. The proposed approach offers a scalable, intelligent, and practical solution for modern cybersecurity environments, supporting organizations in defending against continuously evolving cyber threats while enhancing the effectiveness of threat intelligence systems.
Machine Learning-Based Insider Threat Detection Using CERT Dataset
Authors: Assistant Professor T.Pravalika, B.Gayathri
Abstract: The fact that malevolent actions are carried out by persons with legal access to organisational resources makes insider threats one of the most difficult cybersecurity issues confronted by contemporary organisations. Since suspicious actions typically mimic legitimate user behaviour, insider threats are harder to identify than external intrusions. When dealing with very unbalanced datasets, where harmful events make only a tiny proportion of overall user actions, traditional rule-based security systems often fail to detect modest behavioural anomalies. In order to enhance organisational security via intelligent behavioural analysis, this research proposes a system for insider threat identification that is based on machine learning and uses the CERT 5.2 Insider Threat Dataset. A variety of machine learning algorithms, such as Logistic Regression, Decision Tree, Random Forest, Support Vector Machine (SVM), K-Nearest Neighbours (KNN), Naïve Bayes, AdaBoost, and XGBoost, are included in the suggested framework, along with thorough data preprocessing and the Synthetic Minority Oversampling Technique (SMOTE) to deal with class imbalance. To find out how well the models detect insider threats, we use Accuracy, Precision, Recall, and F1-score. The experimental findings show that among the machine learning models tested, Random Forest and AdaBoost surpass the rest with a classification accuracy of 97.5%, all while retaining outstanding recall and precision. While maintaining the original methodology, dataset, algorithms, and experimental results, the suggested framework offers a scalable and efficient solution for early insider threat detection. This helps organisations with cybersecurity, security risk reduction, and intelligent behavioural analysis for better decision-making.
Machine Learning-Based Hotel Booking Cancellation Prediction And Visual Analytics Framework
Authors: S. Akhila, A. Ruchitha
Abstract: Because they affect occupancy rates, revenue output, and operational planning, reservation cancellations are becoming an increasingly big issue for the hospitality industry. If hotel management can anticipate potential cancellations in advance of the scheduled check-in date, they may put effective reservation strategies in place and make informed decisions. This research presents a machine learning-based system for predicting hotel booking cancellations by assessing reservation-related variables such as booking lead time, guest data, room pricing, market segment, booking history, and special requests. The first step is to apply Exploratory Data Analysis (EDA) on the booking dataset to identify the factors that have a major influence on non-payment. Next, four supervised ML methods are trained and evaluated using various performance metrics, such as F1-score, recall, accuracy, and precision. These algorithms are Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF). When compared to the other models, the Random Forest classifier performs the best in terms of prediction accuracy and total classification performance. Customers will find it much easier to understand the forecast results with the help of an interactive visual analytics tool that shows booking trends, cancellation details, and model comparisons. The proposed solution aids hotel management in a number of ways, including intelligent cancellation prediction, better reservation administration, and data-driven commercial decision-making. The decision-support capabilities of machine learning techniques have also been shown in related disciplines [1, 2].
DOI: http://doi.org/10.5281/zenodo.21374503
An Intelligent Deep Learning Framework For Spoken Language Identification
Authors: K. Rajkumar, A. Swarshitha
Abstract: Modern speech processing applications rely heavily on spoken language identification. This includes intelligent translation systems, voice-based human-computer interaction, and multilingual speech recognition. Communication technologies are enhanced by accurate voice recognition, which allows for efficient processing of multilingual speech data. The five most common Indian languages—Hindi, Bengali, Tamil, English, and Gujarati—are identified using a framework based on deep learning in this study. To classify languages, the suggested technique feeds a Deep Neural Network (DNN) with audio data extracted from speech signals using Mel-Frequency Cepstral Coefficients (MFCCs). To enhance classification performance, audio samples are preprocessed, features are extracted, and models are trained using publically accessible speech datasets. Common performance measures, including as F1-score, recall, accuracy, and precision, are used to assess the efficacy of the suggested model. The results of the experiments show that the deep learning model successfully differentiates between the target languages and shows impressive generalisation when applied to new audio data. In addition, the published web application is built using Django, which incorporates the trained model. Users may access the results of language recognition in real-time by uploading voice recordings and interacting with the interface. With its extensible design, the suggested framework may accommodate more languages and practical speech processing applications while simultaneously providing an efficient, scalable, and workable solution for multilingual spoken language recognition.
DOI: http://doi.org/10.5281/zenodo.21374453
An Intelligent Campus Placement Prediction Framework Using Machine Learning
Authors: S. Srinivas, A. Kalyani
Abstract: Campus placement prediction is a major use of machine learning in the field of higher education. Because of this, schools may gauge a candidate's employability based on their grades and other skill-related attributes. By accurately predicting students' placement results, training and placement cells may better assist students in gaining a better understanding of their strengths and areas for improvement, which in turn helps students in their professional development. This research presents a machine learning-based predictive analytics platform for on-campus placements. The platform uses a variety of criteria, including students' technical ability, internship experience, project work, credentials, and academic records, to forecast how campus placements will go. To eliminate discrepancies, identify significant features, and understand the interrelationships of different variables, the collected student dataset is first subjected to data preparation and exploratory data analysis (EDA). Now we can go on to comparing and applying different supervised machine learning methods. Decision trees, logistic regression, K-nearest neighbours (KNN), and random forests are all examples of such methods. F1-Score, Accuracy, Precision, and Recall are typical performance indicators used in this assessment. Decision Tree and K-Nearest Neighbours classifiers get the highest prediction accuracy when evaluated on the selected dataset. Academic profile analysis and real-time placement predictions are both made possible via an interactive online interface that students and placement managers may utilise. Because of this, the suggested framework is more practically useful. One web program that uses Django is able to use the learned prediction model. This data-driven solution is scalable, effective, and designed to improve campus recruitment strategies, student career guidance, and institutional placement outcomes.
DOI: http://doi.org/10.5281/zenodo.21374561
Deep Learning-Based Gun And Knife Detection Using YOLOv7 For Intelligent Surveillance Systems
Authors: M. Pradeepthi, A. Shravani
Abstract: As the number of crimes committed with firearms and other sharp weapons continues to rise, there is an urgent need for an intelligent monitoring system capable of automatically detecting potential threats. It is not simple to spot knives and guns in surveillance images due to their small size, unpredictable orientations, and complex backgrounds. This research presents a method for the automatic detection of weapons, such as knives and guns, by use of cutting-edge deep learning object recognition algorithms. As its primary detection model, this system compares YOLOv7 against other variations of YOLO, including YOLOv7-x, YOLOv7-w6, YOLOv7-e6, and YOLOv8n. The detection models are evaluated and trained using a tagged dataset consisting of photographs of firearms and knives. The bounding box annotations allow for precise localisation of the weapons inside the input pictures. To assess how well each model performs, metrics including Precision, Recall, mAP@0.5, and mAP@0.5:0.95 are used. Experimental analysis found that the YOLOv7-e6 model had the highest mean Average Precision of the models evaluated, indicating improved detection capabilities and the capacity to more consistently identify weapons like knives and guns in surveillance circumstances. Whether for civilian or military usage, the proposed architecture streamlines the process of intelligent security monitoring and enhances automatic detection of potentially dangerous weapons.
DOI: http://doi.org/0.5281/zenodo.21374697
Intelligent WhatsApp Chat Analysis And Hidden Pattern Visualization Using Sentiment Analysis And NLP
Authors: B. Naresh, A.Vinaya Sri
Abstract: With the explosion of instant messaging apps comes massive volumes of unstructured conversational data, making it harder than ever to spot important patterns of communication and trends in behaviour. One of the most widely used messaging platforms, WhatsApp, stores data on user interactions, facial expressions, and conversation patterns. Still, there are a lot of challenges to solving when analysing this kind of data, like handling casual writing, eliminating unnecessary messages, comprehending sarcasm and contextual meanings, safeguarding user privacy, and efficiently processing massive datasets. In order to find hidden patterns in communication, this research presents a smart WhatsApp conversation analyser that employs sentiment analysis, natural language processing, and interactive data visualisation. To prepare exported WhatsApp chat files for sentiment analysis using the VADER (Valence Aware Dictionary and sEntiment Reasoner) algorithm, the proposed system removes medium messages, unnecessary notifications, and irrelevant information. The system examines sentiment, keyword frequency, user activity, message analytics, and visual exploration using Python modules like Pandas, Matplotlib, and Seaborn. Users are provided with an interactive web application that presents analytical data in order to assist them better understand communication behaviour and emotional tendencies inside talks. Experimental evaluation shows that the proposed framework improves sentiment classification accuracy by around 15% and boosts processing efficiency by more than 20% when compared to conventional conversation analysis methods. The developed approach provides a workable solution to the challenge of extracting actionable information from WhatsApp conversations using machine learning and natural language processing.
DOI: http://doi.org/10.5281/zenodo.21374832
An Intelligent Framework For Early Low Birth Weight Prediction Using Machine Learning
Authors: S. Gouthami, A. Srilekha
Abstract: Low birth weight (LBW), defined as a birth weight of less than 2500 grams, is one of the leading indicators of neonatal morbidity, infant death, and long-term developmental difficulties. Healthcare professionals should be able to detect pregnancies at risk of delivering infants with low birth weights early on, which might improve maternal and newborn outcomes. Traditional risk assessment methods rely heavily on maternal history and clinical findings, but they may not always provide reliable predictions. An artificial intelligence (AI) framework for the early diagnosis and classification of Low Birth Weight is presented in this research, which makes use of demographic, clinical, and prenatal healthcare data for mothers. The proposed system uses a variety of supervised machine learning algorithms for data collection, preprocessing, feature selection, model training, and performance comparison evaluation. These algorithms include Support Vector Classifier (SVC), Decision Tree Classifier (DTC), Random Forest Classifier (RFC), and Extreme Gradient Boosting (XGBoost). The dataset is prepared for training prediction models by performing operations such handling missing values, eliminating outliers, decreasing noise, and optimising features. To determine the most effective prediction model, experimental evaluations are conducted on Accuracy, Precision, Recall, and F1-Score. The Random Forest Classifier has the highest prediction accuracy among the evaluated classifiers, at around 84.81%. A web app developed on Django allows medical professionals to enter maternal health data. The program employs an internal prediction engine to generate real-time risk estimates for low birth weight. The user interface is easy and engaging. A more effective decision-support tool within the proposed framework may lead to better prenatal healthcare planning, earlier diagnosis, and fewer issues with low birth weight in newborns.
DOI: http://doi.org/10.5281/zenodo.21375077
A Study on Investors Preferences on Mutual Funds with Reference to India Bulls Securities Pvt Limited
Authors: E. Prashanth Kumar, Associate Professor Dr. M. P. Suri Ganesh
Abstract: A mutual fund is a programme in which multiple investors pool their funds towards a specific financial goal. The money that was raised was invested in the capital markets, together with the money that was made. The UTI established what was essentially a small savings branch under the RBI, which served as the foundation for the mutual fund industry in India. For the following 25 years, this was reasonably successful since it provided investors with good returns. Due to this, the RBI authorised the establishment of Mutual Funds in India by Public Sector Banks and Financial Institutions. As a result of their success, Private Sector Mutual Funds were able to take off. Portfolio diversification, liquidity, professional management, ease of companies, reduced risk, low transaction costs, transparency, and safety are benefits of mutual funds. It’s incredibly simple to buy and sell mutual funds. Public sector mutual funds and private sector mutual funds are the two categories of mutual funds in India. UTI Mutual Fund, State Bank of India Mutual Funds, and Bank of Baroda Mutual Funds are Public Sector Mutual Funds. Two mutual fund companies, HDFC Mutual Fund & SBI Mutual Fund, have had their returns compared. Both small and midcap companies were included in this comparison. Which markets they invested the investors’ money in and how the returns were calculated over the course of five years. It provides you with suggestions about where and how to make investments. Mutual Funds are Subject to Market Risk; Before Investing, please read the Offer Document.
An Intelligent Weather-Driven Crop Prediction Framework Using Big Data Analytics
Authors: K. Rajkumar, A. Swetha
Abstract: Food security and economic development are two of the most basic human necessities that many countries, India included, depend substantially on agriculture to meet. Variations in temperature, precipitation, humidity, soil type, wind speed, and seasons all have a role in determining harvest success. while farmers don't take these climatic variables into consideration while selecting crops, agricultural production drops and they lose money. This study presents a weather-based crop prediction system that utilises Big Data Analytics and Machine Learning. It has the potential to assist with smart crop recommendation and agricultural decision-making. This proposed system is able to make use of data from large agricultural databases, which include information on crop production, weather, regions, and seasons. By eliminating duplicate or incorrect records, preprocessing ensures that the data is consistent and of high quality. The enhanced dataset is processed using the MapReduce programming paradigm, which allows for efficient handling and analysis of large volumes of agricultural data. After that, we'll use the K-Means Clustering technique to group crops with comparable production characteristics and then figure out the average crop yield in different places. Bar charts and scatter plots are examples of graphical visualisation techniques used to study the relationships between climate conditions and crop production; these tools further enhance understanding of weather-dependent agricultural trends. In addition to assisting farmers in selecting climate-appropriate crops, the developed framework improves agricultural planning and resource utilisation. The system is further implemented via a Django-based web application that provides an interactive platform for weather-based crop suggestion, visualisation, and decision help. Improved crop forecasting and less harmful farming techniques are two outcomes of the proposed method's integration of visual analytics, machine learning, and big data analytics, which contributes to the advancement of precision agriculture.
DOI: http://doi.org/10.5281/zenodo.21375319
Deep Learning-Based Waste Material Classification for Intelligent Automated Trash Segregation
Authors: Assistant Professor P.Premchand Goud, B. soundarya
Abstract: Preserving the environment, promoting sustainable development, and preserving resources all depend on effective waste management. Waste segregation, recycling, and disposal have become more difficult due to the fast growth of municipal solid waste, which has rendered manual classification ineffective and error-prone. To address these issues, this research provides an intelligent waste classification system based on Convolutional Neural Networks (CNNs) for the automated identification of waste elements. Using five predetermined categories—cardboard, glass, metal, paper, and plastic—the suggested approach is able to sort trash photos. Initially, garbage photographs are acquired from a publically accessible dataset and undergo preprocessing processes, including image scaling and augmentation, to increase data consistency and limit the chance of overfitting. The processed pictures are then fed to a CNN architecture consisting of convolutional, activation, pooling, flattening, and fully connected layers for feature extraction and classification. Superbly predicting which trash category an image belongs to, the trained model learns hierarchical visual characteristics automatically from trash photos. The experimental assessment reveals that the suggested CNN-based framework achieves a classification accuracy of 89.88%, with a Precision of 88.12%, Recall of 87.50%, and an F1-score of 87.63%, suggesting its efficacy for automated trash segregation. Furthermore, the trained classification model may be incorporated into an intelligent web-based application to facilitate real-time garbage detection and sustainable waste management methods. The suggested technique helps to enhancing recycling efficiency, eliminating human error, and encouraging environmentally responsible trash disposal using deep learning-based picture categorisation.
A Machine Learning Approach for Real-Time Anomaly Detection in 5G Networks
Authors: Assistant Professor S.Venkateswara Rao, B. Sindhuja
Abstract: With its ultra-high data speeds, low latency, enormous device connectivity, and better service dependability, fifth-generation (5G) cellular networks have greatly improved wireless communication. These networks have been rapidly deployed. But, the attack surface has grown with the ever-increasing complexity of 5G infrastructures and the ever-increasing volume of network traffic, leaving these networks open to cyber attacks, illegal access, and strange communication patterns. The ever-changing nature of 5G settings makes it difficult for traditional security measures to detect complex assaults as they happen. In light of these difficulties, this research introduces a framework for anomaly identification in 5G cellular networks that is based on machine learning. Various machine learning techniques are used by the suggested system. These techniques include Autoencoders, Random Forest, One-Class Support Vector Machine (One-Class SVM), and two ensemble learning methods. Both ensemble models include AdaBoost, Decision Tree, and Gradient Boosting into a Voting Classifier; however, the second model enhances anomaly detection performance by integrating AdaBoost, Decision Tree, and ExtraTree. When creating the models, we used a Cellular Dataset that included both typical and unusual 5G network data. Data cleansing, normalisation, and label encoding are some of the preprocessing activities that the dataset goes through before training in order to make the data better and the model work better. The suggested ensemble strategies are shown to outperform individual machine learning models in extensive experimental assessment. In complicated 5G communication scenarios, Ensemble 2 outperforms all other techniques with 100% Accuracy, Precision, Recall, and F1-score, proving its usefulness for real-time anomaly identification. To improve network security, reduce cyber threats, and enable intelligent monitoring of next-gen cellular communication systems, this study proposes a framework that is both scalable and dependable, all while maintaining the original machine learning methodology.
A Deep Learning Framework For Clinical Facial Skin Disease Identification Using CNN Models
Authors: M.Swathi, B. Ranjana
Abstract: The impact of facial skin problems on one's look and general health makes them a significant dermatological issue. Timely treatment relies on early detection of many illnesses, but manual diagnosis may be time-consuming and frequently requires clinical skill. Convolutional Neural Networks (CNNs) and other deep learning advancements have shown exceptional promise in reliably analysing medical pictures. This study examines the efficacy of convolutional neural network (CNN) models for automated facial skin disease categorisation using clinical face photos. The illnesses under investigation include Rosacea, Actinic Keratosis, Seborrhoeic Keratosis, Lupus Erythematosus, Basal Cell Carcinoma, and Squamous Cell Carcinoma. One of the biggest publicly disclosed collections of face dermatological photos in China, the Xiangya-Derm dataset is used in the experiments. A subset of 2,656 facial photographs was created from this dataset. We took a look at the classification performance of five popular CNN architectures: ResNet-50, Inception-v3, DenseNet121, Xception, and Inception-ResNet-v2. In addition, pretrained models were able to enhance their feature extraction and classification capabilities via the use of transfer learning, which included drawing on an independent dataset that had the same illness categories gathered from other body areas. Experiments show that transfer learning always improves model performance, with Inception-ResNet-v2 doing the best overall. The model achieved recall values of 92.9% for Lupus Erythematosus, 89.2% for Basal Cell Carcinoma, and 84.3% for Seborrhoeic Keratosis on the facial-image test set of 388 clinical images. On average, the model achieved 77.0% recall and 70.8% precision across all six disease categories. The results show that transfer learning has the ability to improve automated skin disease detection in clinical practice and validate that CNN-based facial skin analysis may provide trustworthy assistance for computer-aided dermatological diagnosis.
DOI: http://doi.org/10.5281/zenodo.21375722
A Data-Driven Network Security Situation Awareness Model Using Locality Sensitive Hashing
Authors: G. Preethi, B. Pravalika
Abstract: Network information transmission security has become more complicated due to the ever-increasing complexity of digital communication infrastructures and the ever-increasing volume of network traffic. When dealing with large-scale data settings, traditional methods of network security monitoring may be inefficient and hard to adjust to. In light of these difficulties, this research proposes an LSH-based data-driven paradigm for situational awareness in network security. The suggested architecture analyses and monitors network events in real-time to spot suspicious patterns that might represent security risks. In order to effectively find similar patterns and detect anomalous occurrences, the LSH algorithm is used after data preparation and feature analysis to the network information. To further assess its efficacy in security scenario awareness, the suggested method is compared to the Bayesian algorithm. The LSH algorithm outperforms the Bayesian method, which yields a detection rate of 80% to 85%, according to experimental study, which falls anywhere between 90% and 95%. Furthermore, the LSH-based method consistently identifies anomalies in networks with a fidelity and accuracy that surpasses that of traditional methods, with a false detection rate that remains below 1%. By maintaining the original methodology and experimental evaluation, the proposed framework efficiently improves the security of network information transmission by helping to identify threats faster, increasing situational awareness, and making large-scale communication environments more resilient.
DOI: http://doi.org/10.5281/zenodo.21375863
Deep Learning-Based Waste Material Classification for Intelligent Automated Trash Segregation
Authors: Assistant Professor P.Premchand Goud, B. soundarya
Abstract: Preserving the environment, promoting sustainable development, and preserving resources all depend on effective waste management. Waste segregation, recycling, and disposal have become more difficult due to the fast growth of municipal solid waste, which has rendered manual classification ineffective and error-prone. To address these issues, this research provides an intelligent waste classification system based on Convolutional Neural Networks (CNNs) for the automated identification of waste elements. Using five predetermined categories—cardboard, glass, metal, paper, and plastic—the suggested approach is able to sort trash photos. Initially, garbage photographs are acquired from a publically accessible dataset and undergo preprocessing processes, including image scaling and augmentation, to increase data consistency and limit the chance of overfitting. The processed pictures are then fed to a CNN architecture consisting of convolutional, activation, pooling, flattening, and fully connected layers for feature extraction and classification. Superbly predicting which trash category an image belongs to, the trained model learns hierarchical visual characteristics automatically from trash photos. The experimental assessment reveals that the suggested CNN-based framework achieves a classification accuracy of 89.88%, with a Precision of 88.12%, Recall of 87.50%, and an F1-score of 87.63%, suggesting its efficacy for automated trash segregation. Furthermore, the trained classification model may be incorporated into an intelligent web-based application to facilitate real-time garbage detection and sustainable waste management methods. The suggested technique helps to enhancing recycling efficiency, eliminating human error, and encouraging environmentally responsible trash disposal using deep learning-based picture categorisation.
Advanced Machine Learning Framework For Ocean Wave Energy Prediction
Authors: G. Sudheer Kumar, A. Keerthana
Abstract: Ocean wave energy has great potential for supporting sustainable power production, is continuously available, and has a high energy density, making it one of the most promising renewable energy resources. It is crucial to accurately predict ocean wave properties such as significant wave height (SWH), wave period, and wave energy flow in order to efficiently use wave energy. Prediction accuracy under changing sea conditions is limited by traditional numerical forecasting models' inability to incorporate the ocean's extremely nonlinear, dynamic, and unpredictable character. Through an examination of current developments in ML, DL, ensemble learning, and hybrid prediction models, this research lays forth a framework for intelligent wave energy forecasting that is based on machine learning. The suggested framework takes a look at some of the most popular algorithms for making predictions about the future, such as ANNs, LSTM, CNNs, RF, XGBoost, Transformer-based models, and hybrid ML techniques. These algorithms can predict things like wave height, wave period, and energy generation potential. Hybrid deep learning architectures that include both spatial and temporal feature extraction are more resilient to the unpredictable ocean conditions and provide more accurate forecasts, according to a comparative study. Common metrics used in forecasting for performance assessment include Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Coefficient of Determination (R³). In addition, the suggested forecasting framework is fully integrated into a web application built on Django. This web app lets users analyse oceanographic data, run prediction models using machine learning, see the outcomes of the forecasts visually, and keep an eye on the wave energy production via an interactive dashboard. The suggested system is an intelligent decision-support platform that can optimise wave energy converters, manage renewable energy sources, and integrate ocean wave energy into current power systems efficiently
DOI: http://doi.org/10.5281/zenodo.21376151
A Study on Financial Performance of Maruthi Suzuki
Authors: Athinagaram Abhilash, Dr. S. Narender
Abstract: This study aims to analyze the cash flow statement of Maruti Suzuki India Limited to evaluate its liquidity, operational efficiency, and financial flexibility over a specific period. The cash flow statement, one of the key financial statements, provides insights into the company’s cash inflows and outflows categorized under operating, investing, and financing activities. By examining the cash flow trends, the study helps stakeholders understand how effectively Maruti Suzuki manages its cash resources to sustain operations, invest in growth, and meet financial obligations. The research is conducted using secondary data collected from the company's annual reports and financial statements over the last five financial years. The methodology includes ratio analysis, comparative analysis, and trend analysis to assess the patterns in cash flow movements and their implications on the company’s financial health.
DOI: http://doi.org/10.5281/zenodo.21390732
A Study On Zara’s Global Expansion Strategy and Its Impact On Sales in Emerging Markets With Special Reference to Zara
Authors: B. Vijender Reddy, Dr. Sivaji Jinka
Abstract: Food security and economic development are two of the most basic human necessities that many countries, India included, depend substantially on agriculture to meet. Variations in temperature, precipitation, humidity, soil type, wind speed, and seasons all have a role in determining harvest success. while farmers don't take these climatic variables into consideration while selecting crops, agricultural production drops and they lose money. This study presents a weather-based crop prediction system that utilises Big Data Analytics and Machine Learning. It has the potential to assist with smart crop recommendation and agricultural decision-making. This proposed system is able to make use of data from large agricultural databases, which include information on crop production, weather, regions, and seasons. By eliminating duplicate or incorrect records, preprocessing ensures that the data is consistent and of high quality. The enhanced dataset is processed using the MapReduce programming paradigm, which allows for efficient handling and analysis of large volumes of agricultural data. After that, we'll use the K-Means Clustering technique to group crops with comparable production characteristics and then figure out the average crop yield in different places. Bar charts and scatter plots are examples of graphical visualisation techniques used to study the relationships between climate conditions and crop production; these tools further enhance understanding of weather-dependent agricultural trends. In addition to assisting farmers in selecting climate-appropriate crops, the developed framework improves agricultural planning and resource utilisation. The system is further implemented via a Django-based web application that provides an interactive platform for weather-based crop suggestion, visualisation, and decision help. Improved crop forecasting and less harmful farming techniques are two outcomes of the proposed method's integration of visual analytics, machine learning, and big data analytics, which contributes to the advancement of precision agriculture.
DOI: http://doi.org/10.5281/zenodo.21391884
A Study On Fund Flow Statement With Reference To Zuari Cement Limited
Authors: Parakala Nithin Goud, Dr. S. Narender
Abstract: Finance is the lifeblood of every business enterprise, and the manner in which funds are procured and deployed determines its financial health. Conventional statements — the balance sheet and the profit and loss account — present the position and performance of a business but do not directly reveal the movement of funds between two accounting dates. This article presents a study of the fund flow position of Zuari Cement Limited, a leading South Indian cement manufacturer and part of the Heidelberg Materials group, for the five-year period 2020-21 to 2024-25. The study is based entirely on secondary data drawn from the company's published annual reports, supplemented by journals, textbooks, and authentic websites relating to the cement industry. Adopting an analytical and descriptive research design, the article prepares schedules of changes in working capital, computes funds from operations through the adjusted profit and loss account, and constructs statements of sources and applications of funds, supported by trend analysis and liquidity ratios. The objectives are to examine the conceptual framework of the fund flow statement, analyse changes in working capital, identify the sources and applications of funds, evaluate the efficiency of fund management, and offer suitable suggestions for improvement. The analysis reveals that funds from operations remained positive throughout the study period and constituted the largest source of funds, supplemented by long-term borrowings, while the purchase of fixed assets dominated the applications of funds, reflecting the capital-intensive nature of the cement industry. The net working capital of the company declined in three of the four years, and the liquidity ratios remained below conventional standards, indicating pressure on the short-term financial position.
DOI: http://doi.org/10.5281/zenodo.21408680
A Study On Financial Analysis of Reliance Industires Pvt, Ltd.Hyderabad
Authors: P. Prakash, Dr. k. Pushpa Latha
Abstract: Financial analysis plays a vital role in evaluating the performance, stability, and growth prospects of a company. This study focuses on the financial analysis of Reliance Industries Pvt. Ltd., Hyderabad, with the objective of assessing its financial health using key analytical tools and techniques. The study is based on secondary data collected from annual reports, financial statements, and other reliable financial sources over a specified period. The analysis includes various financial tools such as ratio analysis, trend analysis, and comparative analysis to evaluate profitability, liquidity, solvency, and operational efficiency. Key financial ratios like current ratio, debt-equity ratio, return on capital employed, and net profit margin are used to interpret the company’s performance. Trend analysis helps in understanding the growth pattern of revenue, expenses, and profit over the years. The findings of the study indicate that Reliance Industries has shown strong financial performance with consistent growth in revenue and profitability, supported by efficient resource utilization and diversified business operations. However, certain fluctuations in liquidity and leverage position highlight areas that require strategic attention. The study concludes that financial analysis is an essential tool for decision-making and helps stakeholders such as investors, management, and creditors in evaluating the company’s financial strength and future potential. The research also provides valuable insights for improving financial planning and management practices.
DOI: http://doi.org/10.5281/zenodo.21409762
A Study On Consumer Behaviour and Perception on HDFC Bank
Authors: Mekala Varun, Dr. Sivaji Jinka
Abstract: Consumer behaviour and perception play a vital role in the success of the banking industry. With increasing competition among banks and the rapid growth of digital banking, understanding customers' needs, preferences, and expectations has become essential. This study focuses on analysing the consumer behaviour and perception of customers towards HDFC Bank and the factors that influence their banking decisions. The primary objective of the study is to examine customer preferences regarding HDFC Bank's products and services, evaluate their level of satisfaction, and identify the key factors affecting their perception of the bank. The research is based on both primary and secondary data. Primary data were collected through a structured questionnaire from HDFC Bank customers, while secondary data were gathered from books, journals, research articles, company reports, and the official HDFC Bank website. The study evaluates customer opinions on service quality, digital banking facilities, staff behaviour, accessibility, trust, product variety, complaint handling, and overall banking experience. The findings indicate that most customers have a positive perception of HDFC Bank due to its reliable services, strong digital banking platform, wide range of financial products, and customer-focused approach. However, some customers expect improvements in complaint resolution, personalized services, and waiting time at branches.
DOI: http://doi.org/10.5281/zenodo.21412043
A Study on Cash Flow Statement of Tata Motors
Authors: S. Srikar Reddy, Associate Professor Dr. K. Pushpalatha
Abstract: This study aims to comprehensively analyze the cash flow statement of Tata Motors Limited to evaluate its liquidity, operational efficiency, financial flexibility, and overall solvency over a five-year period (2021-2025). The cash flow statement, one of the key financial statements, provides insights into the company's cash inflows and outflows categorized under operating, investing, and financing activities. For a massive, capital-intensive conglomerate like Tata Motors—which manages the domestic commercial and passenger vehicle market alongside the global Jaguar Land Rover (JLR) luxury brand—cash management is the ultimate determinant of survival.
A Study on Derivatives Future and Options at India Bulls Limited
Authors: Nevuri Praneeth, Associate Professor Mr. Bh.L. Mohanraju
Abstract: The development of the derivatives market particularly instruments such as forwards, futures, and options originated from the need of risk-averse investors to protect themselves against uncertainties caused by fluctuations in asset prices. Derivatives are financial instruments whose value is derived from an underlying asset, and they play a crucial role in risk management. The market primarily consists of three types of participants: hedgers, speculators, and arbitrageurs. Prices in an organized derivatives market reflect the expectations of market participants about future price movements and often influence the prices of the underlying assets. In recent years, derivatives markets have gained significant importance due to their vital role in the financial system. The growth in investments in both domestic and international stock markets has further increased interest in this segment. While extensive research has been conducted in developed markets regarding the impact of futures and options on cash market volatility, the derivatives market in India is still relatively new and not widely understood by all investors. Therefore, regulatory bodies like SEBI need to take proactive measures to create awareness about derivatives trading.
A Comprehensive Study of Smart Manufacturing Using Industry 4.0 Technologies
Authors: Mr. Rahul Ghotkar, Ms. Amisha Malviya, Mr. Rahul Khobragade
Abstract: The manufacturing industry is undergoing a significant transformation driven by rapid advancements in digital technologies, automation, artificial intelligence, and interconnected production systems. Traditional manufacturing methods, which primarily depend on manual operations and isolated production environments, are increasingly being replaced by intelligent, data-driven, and highly automated manufacturing systems. This transformation, commonly referred to as Industry 4.0, integrates advanced technologies such as the Internet of Things (IoT), Artificial Intelligence (AI), Machine Learning (ML), Big Data Analytics, Cloud Computing, Cyber-Physical Systems (CPS), Digital Twins, Robotics, Additive Manufacturing, and Blockchain to create smart manufacturing environments. Smart manufacturing focuses on enhancing production efficiency, reducing operational costs, improving product quality, minimizing resource wastage, and enabling real-time decision-making through intelligent digital systems. As global industries strive for higher productivity and sustainability, Industry 4.0 technologies have become essential components of modern manufacturing strategies. Smart manufacturing represents the evolution of conventional production systems into intelligent and interconnected ecosystems capable of autonomous monitoring, analysis, optimization, and control. Unlike traditional manufacturing, where production planning and quality inspection are often performed manually, smart manufacturing utilizes sensors, intelligent machines, and communication networks to collect real-time operational data throughout the production process. These data are continuously analysed using advanced analytics and artificial intelligence algorithms to identify performance variations, predict equipment failures, optimize production schedules, and improve manufacturing quality. This digital transformation enables organizations to respond rapidly to market demands while maintaining high production efficiency and product reliability. The integration of the Internet of Things (IoT) has become one of the most influential technological developments in Industry 4.0. IoT-enabled sensors continuously monitor machine performance, temperature, pressure, vibration, energy consumption, inventory levels, and production status. The collected information is transmitted through secure communication networks to cloud platforms and manufacturing execution systems for real-time analysis. This continuous flow of information enables manufacturers to monitor production remotely, detect anomalies at an early stage, reduce machine downtime, and improve maintenance planning. Consequently, IoT significantly enhances operational visibility and supports intelligent manufacturing decision-making.
Deep Learning-Based Waste Material Classification for Intelligent Automated Trash Segregation
Authors: Assistant Professor P.Premchand Goud, B. soundarya
Abstract: Preserving the environment, promoting sustainable development, and preserving resources all depend on effective waste management. Waste segregation, recycling, and disposal have become more difficult due to the fast growth of municipal solid waste, which has rendered manual classification ineffective and error-prone. To address these issues, this research provides an intelligent waste classification system based on Convolutional Neural Networks (CNNs) for the automated identification of waste elements. Using five predetermined categories—cardboard, glass, metal, paper, and plastic—the suggested approach is able to sort trash photos. Initially, garbage photographs are acquired from a publically accessible dataset and undergo preprocessing processes, including image scaling and augmentation, to increase data consistency and limit the chance of overfitting. The processed pictures are then fed to a CNN architecture consisting of convolutional, activation, pooling, flattening, and fully connected layers for feature extraction and classification. Superbly predicting which trash category an image belongs to, the trained model learns hierarchical visual characteristics automatically from trash photos. The experimental assessment reveals that the suggested CNN-based framework achieves a classification accuracy of 89.88%, with a Precision of 88.12%, Recall of 87.50%, and an F1-score of 87.63%, suggesting its efficacy for automated trash segregation. Furthermore, the trained classification model may be incorporated into an intelligent web-based application to facilitate real-time garbage detection and sustainable waste management methods. The suggested technique helps to enhancing recycling efficiency, eliminating human error, and encouraging environmentally responsible trash disposal using deep learning-based picture categorisation.
A Machine Learning Approach for Real-Time Anomaly Detection in 5G Networks
Authors: Assistant Professor S.Venkateswara Rao, B. Sindhuja
Abstract: With its ultra-high data speeds, low latency, enormous device connectivity, and better service dependability, fifth-generation (5G) cellular networks have greatly improved wireless communication. These networks have been rapidly deployed. But, the attack surface has grown with the ever-increasing complexity of 5G infrastructures and the ever-increasing volume of network traffic, leaving these networks open to cyber attacks, illegal access, and strange communication patterns. The ever-changing nature of 5G settings makes it difficult for traditional security measures to detect complex assaults as they happen. In light of these difficulties, this research introduces a framework for anomaly identification in 5G cellular networks that is based on machine learning. Various machine learning techniques are used by the suggested system. These techniques include Autoencoders, Random Forest, One-Class Support Vector Machine (One-Class SVM), and two ensemble learning methods. Both ensemble models include AdaBoost, Decision Tree, and Gradient Boosting into a Voting Classifier; however, the second model enhances anomaly detection performance by integrating AdaBoost, Decision Tree, and ExtraTree. When creating the models, we used a Cellular Dataset that included both typical and unusual 5G network data. Data cleansing, normalisation, and label encoding are some of the preprocessing activities that the dataset goes through before training in order to make the data better and the model work better. The suggested ensemble strategies are shown to outperform individual machine learning models in extensive experimental assessment. In complicated 5G communication scenarios, Ensemble 2 outperforms all other techniques with 100% Accuracy, Precision, Recall, and F1-score, proving its usefulness for real-time anomaly identification. To improve network security, reduce cyber threats, and enable intelligent monitoring of next-gen cellular communication systems, this study proposes a framework that is both scalable and dependable, all while maintaining the original machine learning methodology.
Automated People Counting Across Indoor and Outdoor Scenes Through Deep Learning
Authors: Assistant Professor Dr.B.Nageshwar Rao, E. Prathyusha
Abstract: Vision-based people counting has become an essential component of intelligent surveillance systems due to its wide range of applications in public safety, crowd management, smart cities, transportation, retail analytics, and access control. Accurate counting of individuals in both indoor and outdoor environments remains a challenging task because of factors such as occlusion, varying illumination, background clutter, and changes in camera viewpoints. Traditional image processing techniques often fail to provide reliable performance in crowded scenes and dynamic environments. This study presents a deep learning-based vision system for people detection and counting by comparatively evaluating Single Shot Detector (SSD), Faster Region-based Convolutional Neural Network (Faster R-CNN), and multiple versions of You Only Look Once (YOLO), including YOLOv3, YOLOv4, YOLOv5, and YOLOv8. The proposed framework consists of data acquisition, image preprocessing, person detection, centroid tracking, and people counting modules. Experimental analysis is conducted using indoor and outdoor image and video datasets collected under diverse environmental conditions. The comparative evaluation demonstrates that YOLOv8 achieves the best overall performance with an accuracy of 0.92, precision of 0.90, and F1-score of 0.93, outperforming SSD, Faster R-CNN, YOLOv3, YOLOv4, and YOLOv5 in terms of detection accuracy and counting reliability. The proposed framework provides an efficient and scalable solution for real-time people counting while preserving the original methodology, deep learning models, datasets, and experimental findings presented in this study.
Promotion of Micro, Small and Medium Enterprises for The Revival of Rural Economies in Etche Local Government Area, Rivers State, Nigeria
Authors: OGRA Olatunbosun Abiodun Emmanuella, ONYEDIKACHI, Oluchi Jemimah
Abstract: Micro, Small and Medium Enterprises (MSMEs) are widely recognised as important drivers of economic growth, job creation, and sustainable development, especially in rural communities where formal employment opportunities are scarce. This study explored how financial inclusion, government policy support, and technology adoption influence MSME growth and contribute to the revival of rural economies in Etche Local Government Area, Rivers State, Nigeria. A convergent parallel mixed-methods approach was adopted. For the quantitative component, structured questionnaires were administered to 150 MSME operators. The data were analysed using descriptive statistics and multiple linear regression. Complementing this, qualitative insights were gathered through semi-structured interviews with 12 key stakeholders and examined using thematic analysis. The regression results revealed that all three factors had statistically significant positive effects on MSME growth. Technology adoption emerged as the strongest predictor (β = 0.612, p < 0.001), followed by government policy support (β = 0.311, p = 0.047) and financial inclusion (β = 0.287, p = 0.049). Collectively, these variables accounted for 58.4% of the variance in MSME growth (R² = 0.584, Adjusted R² = 0.575, F = 68.42, p < 0.001). The study concludes that technology adoption stands out as the most powerful driver of MSME performance in the study area. It recommends integrated interventions aimed at improving digital infrastructure, expanding financial inclusion, and strengthening policy implementation to foster sustainable rural economic revival in Etche and similar communities across Nigeria.
DOI: http://doi.org/10.5281/zenodo.21638167
Antimicrobial Activity of Adhatoda vasica (Nees) Nees. ex an in Vitro Evaluation of Its Ethanolic Leaf Extract
Authors: Ruchi Upadhya, Varun Jain
Abstract: The escalating global threat of antimicrobial resistance has intensified the need for novel therapeutic agents derived from natural sources. Adhatoda vasica (Family: Acanthaceae), a traditional medicinal plant in Indian systems of medicine (Ayurveda and Unani), is widely used for respiratory and inflammatory disorders. This study evaluated the in vitro antimicrobial activity of ethanolic leaf extracts of A. vasica against common bacterial pathogens. Using the agar disc diffusion method, the extract demonstrated significant inhibition of Staphylococcus aureus (12.6 ± 0.4 mm), Escherichia coli (8.9 ± 0.3 mm), Klebsiella pneumoniae (10.7 ± 0.5 mm), and Candida albicans (13.1 ± 0.2 mm). Minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) were determined via broth dilution, revealing effective concentrations ranging from 62.5–250 µg/mL. Phytochemical screening confirmed the presence of alkaloids (vasicine), flavonoids, and tannins, aligning with reported mechanisms of antimicrobial activity. While the extract outperformed standard antibiotics in some cases, further in vivo studies and compound isolation are warranted to harness its therapeutic potential.