Daily Archives: April 9, 2025

Big Data Analytics for Real-Time Fraud Detection in Insurance Claims
Authors:-Shaba Khatoon , Asst.Prof. Ankita Srivastava, Prof.Shish Ahmad

Abstract-The integration of Artificial Intelligence (AI) and Big Data Analytics is revolutionizing industries by optimizing efficiency, accuracy, and security. In healthcare and insurance, AIdriven Intelligent Document Processing (IDP) automates workflows such as claims automation, medical data extraction, and regulatory compliance management. By utilizing Machine Learning (ML), Natural Language Processing (NLP), and Optical Character Recognition (OCR), IDP accelerates document classification, data validation, and anomaly detection, reducing errors by 90% and cutting processing time by 80%. In the financial sector, AI enhances fraud analytics, risk modeling, and compliance monitoring. Advanced deep learning architectures, pattern recognition, and predictive analytics improve credit risk assessment and real-time fraud mitigation. AI-powered anomaly detection techniques identify suspicious transactions, reducing cybersecurity threats and financial fraud losses.

DOI: 10.61137/ijsret.vol.11.issue2.305

Published by: Kajal Tripathi

Urban Flood Hazard Assessment: Harnessing Ensemble Machine Learning for Next-Generation Risk Analytics
Authors:-Mrs.T.Sankaramma, Ch.Mahesh, M.Venkata Sai Harshith, Shaik Saad, V.Venkata Sai Sanjay, Shaik Mohammad Ashiq Ilahi

Abstract-Urban flood hazard assessment through an ensemble machine learning approach minimizes the bias of individual models and offers a more detailed insight into the evolution of flood risks over time. By integrating diverse models, this approach increases the precision of flood event predictions. In this research, we utilized an ensemble machine learning framework to analyse flood hazards. The results reveal that the ensemble model outperforms conventional methods, such as the classification and regression tree (CART) and random forest (RF). The generated hazard maps confirm the accuracy of the data, facilitating public awareness and pinpointing regions vulnerable to flooding.

DOI: 10.61137/ijsret.vol.11.issue2.304

Published by: Kajal Tripathi

AI-Powered Mental Health Insights: A Comprehensive Review of Machine Learning & Deep Learning Approaches for Social Media Analysis
Authors:-Mrs.R.Veera Meenakshi, B.Vanitha Sri, V.N.V.Karthikeya, P.G.Pranava, K.Uday Meher, A.Sreeja

Abstract-Artificial intelligence is revolutionizing healthcare, particularly in the prediction and diagnosis of various diseases through machine learning (ML) and deep learning (DL) algorithms. With the widespread use of social media platforms like Twitter, Facebook, and Reddit, individuals frequently express their thoughts and emotions online. Mental health has emerged as a significant concern, especially following the COVID-19 pandemic, prompting researchers to leverage ML and DL techniques to analyse social media data for mental health prediction. This study offers a comprehensive review of ML and DL algorithms applied to the prediction of mental disorders, based on an analysis of 37 selected research papers. It presents a comparative accuracy table of ML and DL models for four key mental disorders: Depression, Anxiety, Bipolar Disorder, and ADHD. The findings aim to provide a foundational reference for researchers and practitioners, assisting in future advancements in this field. Additionally, this study compiles a list of publicly available datasets, serving as a valuable resource for further research in mental health analysis using artificial intelligence.

DOI: 10.61137/ijsret.vol.11.issue2.303

Published by: Kajal Tripathi

AI-Driven Global Solar Radiation Prediction: Harnessing Machine Learning and Satellite Imagery for Accurate Forecasting
Authors:-Mrs.A.Srujana Jyothi, M.Siri Sathvika, M.Madhur, I.Chathurya, G.Ram Subhash, K.V.K.Varma

Abstract-Accurate prediction of Daily Global Solar Radiation (DGSR) is crucial for applications in renewable energy, agriculture, and climate studies. This paper explores the effectiveness of Machine Learning (ML) algorithms and satellite imagery in enhancing DGSR prediction accuracy. Traditional ML models typically rely on various meteorological parameters (e.g., temperature, wind speed, atmospheric pressure, and sunshine duration) and radiometric parameters (e.g., aerosol optical thickness, water vapour). In this study, we investigate the impact of incorporating normalized reflectance from satellite images across different spectral channels to improve prediction accuracy. We employ two ML-based regression models: Artificial Neural Networks (ANN) and Support Vector Machines (SVM). The results indicate that the selection of input parameters significantly affects the accuracy of daily solar radiation forecasts. Moreover, the ANN model outperforms SVM, demonstrating superior predictive capability.

DOI: 10.61137/ijsret.vol.11.issue2.302

Published by: Kajal Tripathi

Android Flight Price Prediction Web-Based Platform: Leveraging Generating AI for Real-Time Airfare Forecasting
Authors:-Mrs. M. Mani Deepika, P. Nasivi Ramya Anjani, V. Sai Jyothika Chowdary, Y. Anitha Chowdary, M. Swarna, K.Vamsika

Abstract-The aviation industry faces significant challenges in accurately and swiftly predicting flight fares due to the sector’s dynamic nature. Factors such as fluctuating demand, fuel prices, and route complexities contribute to this unpredictability. To address these issues, this research introduces a novel approach leveraging generative artificial intelligence (GAI) to forecast airfares in real time with high precision. The proposed framework integrates generative models, deep learning architectures, and historical pricing data to enhance predictive accuracy. Utilizing GAI within an advanced web engineering framework, this method effectively captures intricate patterns and relationships within historical airline data. By employing deep neural networks, the model efficiently processes diverse scenarios, extracting critical insights to improve the understanding of key factors influencing flight costs. Furthermore, the approach prioritizes real-time forecasting, enabling rapid adaptation to market fluctuations and providing valuable insights for dynamic pricing strategies.

DOI: 10.61137/ijsret.vol.11.issue2.301

Published by: Kajal Tripathi

Intelligent Railway Track Fault Detection Using Image Processing and Fuzzy Logic for Enhanced Safety
Authors:-Mrs. G.Tejasri Devi, P.H. Naga Datta Sanjeev, A.Kasi Viswanadh, A.Sankar, P.Veera Mahesh, Y.Lakshmi Chakradhar

Abstract-The advancement of railway transportation vehicles significantly affects the transportation network. Various errors occur due to the utilization of train lines, arising from both manufacturing defects and improper rail usage. Early detection and correction of these faults are crucial, and several techniques have been developed to address this issue. One effective method involves the use of camera-based systems. By employing cameras mounted on railway vehicles, images of rail components are captured and analysed to identify potential defects. This paper proposes a method for detecting and classifying defects on rail track surfaces using image processing techniques. The system relies on high-resolution images obtained from specialized cameras installed on railway inspection vehicles. These images are analysed to identify and assess various track anomalies, including cracks, welding defects, track misalignment, and ballast deterioration. The image processing workflow involves pre-processing, feature extraction, and segmentation to isolate the rail area and detect potential faults. To prioritize maintenance activities, fuzzy logic is applied after identifying and evaluating the severity of defects. This approach is particularly effective in handling the uncertainty and imprecision associated with track condition assessments. Fuzzy rules and membership functions are designed to assign severity levels to the extracted features of each defect category. This method offers a comprehensive and adaptable solution for improving railway track maintenance and ensuring operational safety.

DOI: 10.61137/ijsret.vol.11.issue2.300

Published by: Kajal Tripathi

Next-Gen Gait Recognition: Advanced Machine Learning for Precision Biometric Analysis
Authors:-Mr.Y.Ravi Bhushan, K.Charan Praveen Kumar, M.Sushma, T.Lasya Srivallika, Ch.Geetha Sri, K.D.V.Chaitanya

Abstract-Stroke, which ranks as the second leading cause of death worldwide, requires prompt and precise prediction for effective intervention. This study conducts a comprehensive exploration of gait recognition in biometric analysis, addressing the unique challenges of using gait as an identifier. It systematically evaluates various machine learning algorithms, including Individual Node Evaluation, Statistical Inference, Regression Modelling, Support Vector Machines, k-Nearest Neighbours, Decision Trees, Random Forests, and Neural Networks. Each model undergoes rigorous testing to assess its effectiveness in accurately identifying individuals based on their gait patterns. The methodology emphasizes thorough preprocessing to maintain data integrity and relevance, incorporating Sequential Backward Selection (SBS) for feature selection and dimensionality reduction techniques like Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) to enhance model efficiency. Additionally, the study explores deep learning architectures, analysing their impact on recognition accuracy. A detailed comparative analysis highlights the strengths and weaknesses of each approach, offering valuable insights into the field. By evaluating a range of ML and DL techniques, this research sets a benchmark for future advancements in biometric security, reinforcing gait recognition as a reliable, non-invasive identification method and paving the way for advanced biometric systems in security and personal identification.

DOI: 10.61137/ijsret.vol.11.issue2.299

Published by: Kajal Tripathi

Smart Stroke Detection: Cutting-Edge Machine Learning and Optimized Algorithms for Early Diagnosis
Authors:-Mr. N.V.S Gopalam, K.Tanoosh, Ch.Sowjanya, Y.Navatej, K.Banny, B.Lakshmi Jahnavi

Abstract-Stroke, which ranks as the second leading cause of death worldwide, requires prompt and precise prediction for effective intervention. This research investigates the use of advanced machine learning techniques to improve stroke prediction models. Initially, classifiers such as Decision Tree (DT), Random Forest (RF), and Support Vector Machine (SVM) were applied, followed by the incorporation of advanced algorithms like Extreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LGBM) to enhance predictive accuracy. Various evaluation metrics, including accuracy, sensitivity, error rates, and log loss, were employed to assess the performance of the models. The findings demonstrate the effectiveness of machine learning algorithms, with XGBoost achieving an impressive accuracy rate of 98%. Additionally, LGBM played a significant role in boosting overall accuracy. These results highlight the critical contribution of advanced machine learning techniques to enhancing stroke prediction. By leveraging these state-of-the-art predictive models, the study advocates for their integration into clinical settings, aiming to expedite accurate diagnoses, improve patient care, and advance stroke detection capabilities. Keywords: Brain Stroke, Machine Learning, Light Gradient Boosting Machine (LGBM), Extreme Gradient Boosting (XGBoost).

DOI: 10.61137/ijsret.vol.11.issue2.298

Published by: Kajal Tripathi

Smart Parking Management Assessment Using Machine Learning Algorithms and IOT
Authors:-Assistant Professor Meenakshi Thalor, Ishwari Abuj

Abstract-With the growing number of vehicles and limited parking infrastructure, parking space man emerged as a major challenge in urban areas. In this paper, an extensive study of machine l models in an IoT-supported space is given, focusing on proposing an ML-based model that available parking space. The study compares the performance of several models Typed as (KNNs), support vector machines (SVMs), random forest (RF), decision tree (DT), logistic and Naive Bayes (NB) regarding to “precision, recall, accuracy, and F1-score performance results obtained after running ML models on the data with 65% and 85% threshold are com meaningful insights about their efficiency of prediction in parking vacancy. Random Forest (RF) model shows the best performance based on those metrics in all evalu high precision, recall, accuracy and F1-score values. The IoT-enabled environment shows t showing its effectiveness in falsely predicting parking space availability. In contrast, K- ne (KNNs), decision tree (DT), logistic regression (LR), predicting Naive Bayes (NB) with co exhibit relatively lower performance in crowded parking GLES scenarios. The paper ends deployment of intelligent predictive models, especially random forest, improves substantial and performance of smart parking system as well as it frees waiting time for cars, and henc parking resource utility as well as it decreases real-time travel congestion and increases use environments.

DOI: 10.61137/ijsret.vol.11.issue2.297

Published by: Kajal Tripathi

Miscellaneous Trends in it
Authors:- Yuvraj Lolage, Sara Lonare, Aditi Londhe, Mrs. Anuja S. Phapale

Abstract-The vast, ever-shifting landscape of human innovation information technology (IT) stands as both a mirror and a catalyst of our collective aspirations. Information technology repeatedly shapes our modern world, exerting influence upon government, industries and daily life. Beyond the headline-grabbing revolutions of artificial intelligence, cloud computing, and blockchain lie quieter, yet equally transformative, currents of change. As humanity ventures further into the digital age, it becomes clear that technology is not merely a tool; it is a partner in shaping the narrative of progress and a testament to the boundless curiosity that drives us to explore the unknown. This paper seeks to delve into these emerging trends, exploring their technical gradation and their broader implications for society. By analyzing their significance, potential applications, and implications for the future, the study aims to provide a comprehensive understanding of how these emerging trends are influencing the broader IT domain.

DOI: 10.61137/ijsret.vol.11.issue2.296

Published by: Kajal Tripathi