Authors: Vikram S Tigadi, Yallaling R Dalawayi, Rajesh S Meti,, Rajguru M Hiremath, Professor Pooja C Shindhe

Abstract: Heart disease remains one of the leading causes of death throughout the world, and early detection is the key to improved patient outcomes. This paper introduces a Decision Support Heart Health Prediction System (DSHHPS) developed using machine learning techniques to help diagnose critical clinical and demographical data including age, BP level, cholesterol level, glucose level and other vital medical signs. The processed data is further sanitized using pre-cleaning, preprocessing and selection of features to make it reliable and accurate. Several different machine learning models are tested and compared The system evaluates many clinical information such as age, sex, blood pressure, cholesterol level, the results of the resting ECG reading, the type of chest pain and the amount of sugar in their bloodstream along with other important health readings. Rigor: The dataset is subjected to various cleaning, preprocessing and feature selection processes to remove inconsistencies and error prior to training the model. A number of machine learning models are experimented and compared to select the best one, which produces the most accurate predictions.

Published by: vikaspatanker

Authors: Wafaa N. Jasim, Faten N. Jasim

Abstract: To characterize and analyze the capability of elements to attenuate X-rays, a number of important physical indicators were calculated, namely the mass attenuation coefficient µ/ρ, the half-value layer HVL, and the free path rate MFP, as they play a role in describing the nature of the material that can be used as a protection method in medical centers and laboratories related to dealing with X-rays, using two methods. The first is using the XCOM program, and the second method is the FFAST tool. They were applied to the elements Fe, Ag, Sn, Pt, Au, and Pb at an energy of 30 KeV, in order to evaluate the effectiveness of each element in medical and industrial applications related to radiation protection. We obtained good agreement between the two methods, and the Matlab program was relied upon in the calculations and drawings that show the relationship and agreement between them.

DOI: https://doi.org/10.5281/zenodo.17949497

Published by: vikaspatanker

Authors: Dr.C.K Gomathy, Ananth Lakshmi ss, Lakshmi A

Abstract: Urban areas are becoming increasingly congested as populations grow and public spaces experience unpredictable fluctuations in foot traffic. This constant movement creates challenges for city planners, traffic authorities, and public safety teams who require reliable, real-time information to manage crowds efficiently. This research investigates the use of Big Data Analytics for predicting urban crowd flow by analyzing digital footprint signals generated through everyday human interaction with technology. These signals include smartphone GPS activity, Wi-Fi hotspot connections, public sensor logs, transport card swipes, and metadata from CCTV systems. By integrating these diverse and continuous data streams, the study proposes a multi-layered predictive framework capable of detecting mobility patterns, forecasting future crowd density, and supporting city-level decision-making. Through machine learning and deep learning models, the framework processes large-scale movement data and produces highly accurate predictions. The findings demonstrate that Big Data-driven analysis significantly enhances crowd-flow forecasting accuracy, improves safety management, supports effective traffic control, and strengthens urban planning strategies for smart cities.

DOI:

Published by: vikaspatanker

Authors: Dr Hari Gangadhar Kale

Abstract: Many aspects of life have benefited from laser technology which is regarded as one of the most significant technologies of the 20th century. These days laser technology is valuable in all manufacturing fields and offers a number of unique advantages including the production of mechanical tools and machines. Laser technology has steadily taken over and dominated the mechanical market particularly in the areas of material handling and metal parts because of its advanced cleaning capabilities fine welding lines powerful etching strokes high power operation and precise distance measurement capability. The benefits this business receives from laser cutting technology.

Published by: vikaspatanker

Authors: Sharmila P, Harini S, Harini D, Harini RV, Jothivarshini S, Karthi S

Abstract: In recent years, personalized fashion recommendation systems have gained significant importance due to the growing demand for customized user experiences in online shopping platforms. This project presents a Smart Dress Recommendation System that provides personalized clothing suggestions based on individual user preferences and physical attributes. The system employs a rule-based recommendation approach, where users are guided through a structured questionnaire to collect essential information such as occasion, gender, budget, style preferences, body measurements, body shape, and skin tone. The body measurement module analyzes user inputs to classify body types such as pear, rectangle, apple, or hourglass, while the skin tone module allows users to select their skin color from a predefined palette. Based on these parameters, a set of predefined rules is applied to recommend suitable dresses that enhance the user’s appearance and meet their personal preferences. The system ensures that fashion suggestions are practical, affordable, and visually appealing by considering budget constraints and occasion-specific requirements.

 

Published by: vikaspatanker

Authors: Srinivasa Chakravarthy Seethala

Abstract: This research paper investigates the role of machine learning enhanced query optimization techniques in improving the performance, scalability, and reliability of SQL and NoSQL databases used for high volume big data processing in financial and healthcare applications. Traditional rule based and cost based query optimizers often struggle to adapt to dynamic workloads, heterogeneous data distributions, and rapidly evolving access patterns that characterize modern financial transactions and healthcare data ecosystems. This study addresses the research problem of how adaptive machine learning driven optimization models can overcome these limitations by learning from historical query execution patterns, system telemetry, and workload characteristics. A mixed methodology is adopted, combining conceptual framework design, algorithmic modeling, and comparative performance analysis across representative SQL and NoSQL environments handling large scale transactional and analytical workloads. The findings demonstrate that machine learning enhanced optimizers significantly reduce query latency, improve resource utilization efficiency, and enhance workload predictability under peak data volumes compared to traditional optimization approaches. The paper highlights key innovations such as predictive cost modeling, adaptive index selection, and real time execution plan refinement driven by supervised and reinforcement learning techniques. From an academic perspective, this research contributes to the evolving discourse on intelligent data management systems by extending optimization theory into data driven adaptive architectures. From an industry standpoint, the results provide actionable insights for designing resilient, high performance data platforms capable of supporting mission critical financial and healthcare operations where accuracy, compliance, and responsiveness are paramount.

DOI: http://doi.org/10.5281/zenodo.17938651

Published by: vikaspatanker

Authors: Srinivasa Chakravarthy Seethala

Abstract: This study proposes an AI driven compliance intelligence platform designed to enable continuous monitoring and automated risk assessment within highly regulated CRM and ERP environments. Organizations operating in finance, healthcare, public sector, and other compliance intensive domains increasingly rely on complex enterprise platforms where regulatory obligations evolve faster than traditional audit and control mechanisms can adapt. The research addresses the limitations of static compliance models by introducing an architecture that integrates machine learning, natural language processing, and policy aware analytics to interpret regulatory requirements, monitor transactional and configuration level signals, and dynamically assess compliance risk in real time. A mixed methodological approach is adopted, combining conceptual system design with simulated enterprise data flows and scenario based evaluations across common regulatory regimes such as data protection, financial controls, and access governance. The findings demonstrate that AI driven compliance intelligence can significantly improve early risk detection, reduce manual audit effort, and enhance traceability across CRM and ERP processes by continuously correlating system behavior with regulatory intent. The platform introduces adaptive risk scoring, automated control validation, and explainable compliance insights that support both operational teams and governance stakeholders. From a strategic perspective, the study contributes to a forward looking compliance paradigm that shifts organizations from periodic, reactive audits toward proactive and continuous assurance models. Academically, the research extends existing literature on enterprise governance by formalizing compliance intelligence as a scalable, data driven capability embedded within enterprise software ecosystems.

DOI: http://doi.org/10.5281/zenodo.17938605

Published by: vikaspatanker

Authors: Srinivasa Chakravarthy Seethala

Abstract: Real time event streaming pipelines form the operational backbone of modern digital platforms, supporting continuous data ingestion, processing, and delivery across cloud native and distributed environments. Despite their importance, reliability management in such pipelines remains largely reactive, relying on threshold based monitoring and post failure diagnostics that are insufficient for preventing cascading disruptions. This study addresses the problem of anticipating reliability degradation in real time event streaming systems by proposing a predictive reliability engineering framework grounded in multi modal deep learning. The primary objective is to enable early identification of failure precursors by jointly analyzing heterogeneous telemetry sources, including system metrics, execution logs, distributed traces, and event level metadata. A mixed method research approach is adopted, combining quantitative modeling of historical incident data with qualitative architectural analysis of streaming platforms to inform model design and integration. The proposed framework employs temporal and representation learning techniques to fuse multi modal signals and generate probabilistic reliability risk scores ahead of observable failures. Experimental evaluation across representative streaming workloads demonstrates improved failure prediction accuracy, longer warning lead times, and reduced false alert rates compared to single source monitoring baselines. The findings highlight the innovation of multi modal fusion for reliability prediction and its implications for proactive operational decision making. From an academic perspective, the study advances reliability engineering by introducing predictive, data driven models tailored to real time pipelines. From an industry standpoint, the framework supports more resilient event driven architectures through earlier intervention, reduced downtime, and improved service continuity, reinforcing the strategic value of intelligent reliability management in high availability systems.

DOI: http://doi.org/10.5281/zenodo.17938555

Published by: vikaspatanker

Authors: Mridulaxika, Gurpreet Singh, Varuna Tyagi

Abstract: Student dropout is a persistent challenge in higher education, leading to academic, financial, and institutional losses. Accurate early prediction of at-risk students can significantly improve retention through timely interventions. This paper presents a comparative analysis of three ensemble-based machine learning models AdaBoost, Gradient Boosting Machine (GBM), and a proposed Advanced Extreme Gradient Boosting (ADVXGBoost) algorithm for predicting student dropout risk. The models were evaluated using a dataset of 5,000 student records containing demographic, academic, and behavioral attributes. Performance was assessed using 10-fold stratified cross-validation in the WEKA Explorer environment. Experimental results demonstrate that ADVXGBoost outperforms AdaBoost and GBM, achieving the highest accuracy of 90.76%, the lowest error rates, and balanced class-wise prediction. The findings confirm the effectiveness of enhanced boosting techniques for reliable student dropout prediction and decision-support systems in educational institutions.

Published by: vikaspatanker

Authors: Dr. Arati Basumatary

Abstract: Social media has become a special part of society. In the present age, social media is considered the best medium for communication across the world. The availability of the internet has made social media popular with the facility of instant exchange. Especially platforms like Facebook, WhatsApp, Instagram, YouTube etc. have facilitated communication, photo-video sharing from anywhere. From the new generation to the older generation, people are now attracted to and experienced with the use of such social media. It can be said that social media is a suitable platform not only for entertainment but also for education, business etc. The widespread use and popularity of social media is also observed in Assamese culture. Assamese songs-dances, food etc, the original cultural elements have been able to be presented on the world stage through social media. However, in this context, it can be assumed that the authenticity and values of Assamese traditional culture are somewhat hindered. This proposed research paper will also discuss the impact of social media on Assamese culture, including both positive and some negative impacts.

Published by: vikaspatanker