Stroke is still a significant global health concern that requires sophisticated predictive methods for early detection and treatment. In this work, a novel machine learning (ML) framework for automated stroke prediction is presented, and its accuracy and generalization skills are evaluated against those of six well-known classifiers. In order to guarantee clear decision-making in clinical applications, SHAP and LIME approaches are also used to highlight model interpretability. By combining local and global analytical approaches, the suggested framework improves the standardization of intricate machine learning models. Notably, Random Forest routinely achieves higher predicting accuracy than other algorithms. An enhanced ensemble strategy that uses a voting mechanism to leverage numerous classifiers and incorporates CATBOOST and a Stacking Classifier is presented in order to further increase performance. This study offers a thorough and trustworthy approach to early stroke diagnosis and treatment, which will ultimately lessen the serious health and financial effects of this common illness.

Keywords : Data Integrity, Duplication detection, Data Validation, Anomaly Detection, Automated alerts.

Authors : BOKKA SURI BABU

Title : Enhancing Brain Stroke Prediction Using Machine Learning for Early Intervention

Volume/Issue : 2025;7(1 (January - February))

Page No : 46 - 52

The key challenges in solving big data units are the integrity and duplication of facts, as mistakes and layoffs can easily degrade choice -making and operational efficiency. This venture applies to the development of sturdy statistics integrity and duplicate caution system, which ambitions to discover, confirm and alleviate statistics and transmissions throughout exclusive facts assets. The procedure integrates records ingestion, validation and determination to preserve the individuality of records. For established and unstructured information, it makes use of a solution primarily based on rules, hash -primarily based determination and anomaly detection with advanced strategies for efficient processing. In addition, the mixing of the warning mechanism guarantees instantaneous bulletins of viable troubles that allow on the spot remedies. The implementation of audit protocols ensures the traceability and duty in fixing the detected anomalies and their possible resolutions. The scalp and flexible system helps integration with exceptional databases.

Keywords : Data Integrity, Duplication detection, Data Validation, Anomaly Detection, Automated alerts.

Authors : Dillibabu

Title : Data Integrity and Deplication Alert System

Volume/Issue : 2025;7(1 (January - February))

Page No : 33 - 38

The Intelligent Chatbot Ticketing System leverages advanced artificial intelligence techniques, particularly Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN), to automate and enhance the process of managing and responding to customer service tickets. This system aims to streamline the ticketing process by efficiently understanding and categorizing user inquiries, providing quick resolutions, and improving overall user experience. The RNN is employed to capture sequential patterns in user interactions, enabling the system to understand context and provide coherent, context-aware responses. Meanwhile, CNN is used for feature extraction from text inputs, helping the system identify important keywords and topics related to customer queries. This combination of RNN and CNN allows the system to not only classify and prioritize tickets effectively but also learn from past interactions to continuously improve its performance. The integration of these deep learning models ensures an intelligent, adaptive, and user-friendly ticketing system that reduces the manual effort required from customer service agents and provides timely assistance to users.

Keywords : Neural Networks, CNN, AI, Customer Service Automation, E-commerce Support.

Authors : Tarollu prasad

Title : Intelligent Chatbot Ticketing System Using RNN and CNN

Volume/Issue : 2025;7(1 (January - February))

Page No : 30 - 33

Blockchain technology with its promise of decentralized, transparent, and tamper-proof systems, it unquestionably captured the imagination of the technology industry and beyond. Blockchain technology has emerged as a revolutionary force that has the capacity to transform industries by enhancing security, ensuring data integrity, and enabling decentralized transactions. Regardless of its promising benefits, blockchain faces critical difficulties that need addressing to open its maximum capacity. In this paper we introduced tied down Information management, key moves and blueprints nitty gritty techniques to beat them.

Keywords : Finney Attack, Race Attack, Time Jacking Attack, Data Management, Security.

Authors : Premkumar

Title : Activity of Blockchain Technology in Secure Data Management and Strategies used to Forestall Attacks

Volume/Issue : 2025;7(1 (January - February))

Page No : 25 - 29

Fraud has come a trillion-bone assiduity moment. Certain fiscal institutions have devoted brigades of sphere experts and data scientists that are assigned with detecting fraudulent exertion. To find frauds, data scientists constantly employ intricate statistical models. But there are a lot of downsides to this strategy. Since fraud discovery isn't real- time, fraudulent exertion are constantly discovered only after the factual fraud has taken place. These methodologies are prone to mortal crimes. In addition, it requires precious, largely professed sphere expert brigades and data scientists. nonetheless, the delicacy of homemade fraud discovery methodologies is low and due to that, it's veritably delicate to handle large volumes of data. More frequently, it requires time- consuming examinations into the other deals related to the fraudulent exertion in order to identify fraudulent exertion patterns.Fraud has come a trillion-bone assiduity moment. Certain fiscal institutions have devoted brigades of sphere experts and data scientists that are assigned with detecting fraudulent exertion. To find frauds, data scientists constantly employ intricate statistical models. But there are a lot of downsides to this strategy. Since fraud discovery isn't real- time, fraudulent exertion are constantly discovered only after the factual fraud has taken place. These methodologies are prone to mortal crimes. In addition, it requires precious, largely professed sphere expert brigades and data scientists. nonetheless, the delicacy of homemade fraud discovery methodologies is low and due to that, it's veritably delicate to handle large volumes of data. More frequently, it requires time- consuming examinations into the other deals related to the fraudulent exertion in order to identify fraudulent exertion patterns.

Keywords : Random Forest Classifier, Decision Tree Classifier, K-Means, Gaussian NB, SVC and K Neighbors Classifier.

Authors : Lakshmiprasanna

Title : A Predictive Analytics with Machine Learning Algorithms for Real Time Financial Data

Volume/Issue : 2025;7(1 (January - February))

Page No : 21 - 24

The early identification of Down Syndrome serves as a necessity to obtain proper medical care together with support services. Actual diagnostic systems use karyotyping along with genetic testing yet these techniques remain expensive while being intrusive and taking much time to complete. This research develops a new transfer learning-based method which diagnoses DS through facial image assessment without causing harm to patients. The VNL-Net system merges three components to perform diagnosis: it starts with spatial feature extraction through VGG16 before utilizing Non-Negative Matrix Factorization for dimensionality reduction then finishes with Light Gradient Boosting Machine for feature generation. K-fold cross-validation combined with a Logistic Regression classifier provides the system with final dataset classification capabilities through performance reliability across different datasets. The researchers created a MobileNet + SVM hybrid model as a solution for efficient diagnosis which enables operation on mobile devices and edge platforms. The facial image data for training the system comes from the publicly accessible Kaggle dataset that shows pictures of both DS-affected children along with unaffected ones. The new system demonstrates its effectiveness through extensive testing that confirms it obtains superior diagnostic performance compared to traditional machine learning algorithms. Hybrid classification and transfer learning work together to achieve better diagnosis results while requiring minimal hardware resources. This study establishes deep learning as a promising technology for medical image analytic applications which provides a cost-efficient tool for diagnosing Down Syndrome early. The proposed project seeks to enlarge the database collection and integrate explainable AI systems to enhance clinical acceptance through improved interpretability.

Keywords : Medical Image Analysis, Logistic Regression, Mobile Net, SVM.

Authors : Vijaya Bhaskar Reddy

Title : A Transfer Learning Approach for Down Syndrome Detection Using Facial Image Features

Volume/Issue : 2025;7(1 (January - February))

Page No : 15 - 20

The Mango leaf diseases significantly restrain mango output as they affect yield and tree health. Mango leaf sooty mould disease is one of the diseases which affects the tree’s photosynthesis and general vigor quite severely. This research study looks into the use of deep learning models like the Residual Network with 50 stacks and ResNext50 for assessing that severity classification of mango leaf sooty mould disease. The model evaluates severity based on the dataset of 25,000 images obtained from different mango fields, as per the study. The overall accuracy achieved is 94.61% for the ResNext50 architecture through layer-wise parameter analysis, performance metrics, and confusion matrices. Model comparisons in the field reveal advantages across and between models. This research not only proves the use of DL in disease management but also paves the way for more applications in farming use. Automated mango leaf disease assessment is always bright white.

Keywords : Machine Learning, Precision Agriculture, Image Recognition, Crop Management, Environmental Data.

Authors : N Divya

Title : AI driven Mango Plant Disease Preduction and Management System

Volume/Issue : 2025;7(1 (January - February))

Page No : 9 - 14

Efficient waste management is crucial in reducing environmental pollution and promoting sustainability. This paper presents an automated waste classification and sorting system that integrates deep learning with Arduino-based hardware for real-time waste disposal. The system employs the YOLOv8 object detection model to classify waste into three categories: biodegradable, non-biodegradable, and hazardous. A laptop's integrated camera captures waste images, which are processed by a deep learning model to determine the appropriate bin for disposal. The hardware system consists of an Arduino board controlling servo motors that open designated pressable bins based on the classification. Experimental results demonstrate high classification accuracy and efficient real-time sorting, reducing human effort and enhancing waste management efficiency.

Keywords : Waste Classification, Deep Learning, YOLOv8, Arduino, Object Detection, Automated Sorting.

Authors : Deepak

Title : Automated Waste Classification and Sorting System Using Deep Learning and Arduino

Volume/Issue : 2025;7(1 (January - February))

Page No : 1 - 8