Category Archives: Uncategorized

Authors: Chetan Uddhav Wagh, Yash Tatyaba Khandare, Shreya Prashant Mhaskar, Prof. K. R. Metha

Abstract: Artificial Intelligence based conversational systems are transforming how information services are delivered in many fields including education. Students frequently require quick access to academic information such as course details, admission procedures, faculty information, and examination schedules. However, traditional communication channels such as notice boards, websites, and help desks often fail to provide instant responses. This research proposes ChatGMVIT, an AI-powered chatbot designed to assist students by providing instant responses to institutional queries. The system is implemented as an Android application integrated with Firebase Firestore as a knowledge database and the Gemini 2.5 Flash Lite model for natural language processing.

DOI: https://doi.org/10.5281/zenodo.19593570

Published by: vikaspatanker

Authors: Mr.B. Janu Naik, Hemadri Sumedha, Anala Sanghavi, Giduthuri Charanchandu, Sathi Sanjana Reddy, Geetha Yeswanth Kumar

Abstract: Forecasting financial market trends has become an important task for investors, financial institutions, and analysts due to the increasing complexity and volatility of modern financial systems. Accurate prediction of market movements such as stock prices, exchange rates, and commodity prices can significantly assist in making informed investment decisions and managing financial risks. However, financial market data is highly dynamic, nonlinear, and influenced by multiple economic and external factors, which makes accurate forecasting a challenging problem for traditional statistical methods. In this study, a machine learning-based framework is proposed for forecasting financial market trends using time series analysis. The proposed approach utilizes historical financial data including stock prices, trading volumes, and other relevant financial indicators to train predictive models capable of identifying patterns and relationships within the data. A Support Vector Machine (SVM) algorithm is employed as the primary forecasting model due to its strong capability to handle nonlinear relationships and high-dimensional datasets effectively. The system performs several important stages including data loading, preprocessing, feature selection, model training, and performance evaluation. During preprocessing, missing values and irregularities in the dataset are handled, and normalization techniques are applied to ensure consistent feature scales. The processed data is then used to train the SVM model, which learns complex patterns present in historical financial data to generate predictions for future market trends. The performance of the forecasting model is evaluated using standard evaluation metrics such as Mean Squared Error (MSE) and Mean Absolute Error (MAE) to measure prediction accuracy. Experimental analysis demonstrates that the proposed machine learning framework is capable of effectively capturing nonlinear patterns present in financial time series data and providing reliable forecasting results. By leveraging machine learning techniques, the proposed system improves prediction efficiency and supports intelligent decision-making for traders, investors, and financial analysts. The framework can serve as a valuable tool for financial market analysis and can be further enhanced by integrating hybrid machine learning models and real-time financial data sources.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.180

Published by: vikaspatanker

Authors: Ms. Y Suma Chamundeswari, Nalli Neeharika, Sneha Dindi, Abbireddy Durga Devi, Nyasavarajula R S Gowtham Datta, Ayanamahanthi Thandava Krishna Murthy

Abstract: Hearing impairment is one of the most common sensory disorders affecting newborns, infants, and young children worldwide. Early detection of hearing loss is crucial because delayed diagnosis can negatively affect speech development, cognitive growth, social interaction, and educational outcomes in children. However, many developing and underdeveloped regions face a shortage of audiologists and otolaryngologists, which often results in delayed diagnosis and limited access to hearing care services. This situation highlights the need for automated and intelligent diagnostic tools that can assist healthcare professionals in identifying hearing impairments more efficiently. This study proposes an automated hearing loss detection framework based on machine learning techniques to support medical professionals in diagnosing hearing impairments in newborns, infants, and toddlers. The proposed system integrates a hearing test data generation module with a machine learning classification model capable of analyzing audiometry test data and predicting the presence and characteristics of hearing loss. The data generation module creates a comprehensive dataset representing different hearing conditions, which is then used to train and evaluate the machine learning model. By employing multiclass and multi-label classification techniques, the model can identify the type, degree, and configuration of hearing loss with high accuracy. Experimental results demonstrate strong diagnostic performance, achieving a prediction time of approximately 634 milliseconds, a log-loss reduction rate of 98.48%, and macro and micro precision values close to 100%. These results indicate that the proposed framework can provide rapid and reliable diagnostic support for healthcare professionals, enabling earlier intervention and improving access to hearing care in regions with limited medical resources.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.181

Published by: vikaspatanker

Authors: Mr .B.Janu Naik, Velugubanti Lakshmana Siva Ganesh, Vignesh Mullangi, Vanapalli Veera Satya Sai Praneesh, Maddula Veera Venkata Sai Pradeep

Abstract: With the rapid expansion of Android applications, malware attacks targeting mobile devices have increased significantly, creating serious security and privacy concerns for users. Traditional malware detection approaches, such as signature-based and rule-based methods, often fail to detect newly emerging or obfuscated malware variants. To overcome these challenges, this study proposes an explainable artificial intelligence-based framework, named XAI-Droid, for effective Android malware detection and classification. The proposed system integrates deep learning techniques with explainable AI (XAI) methods to enhance detection accuracy while ensuring transparency and interpretability in decision-making. Feature extraction is carried out using static analysis techniques, and the extracted features are used to train advanced machine learning and deep learning models. To improve trust and reliability, explanation methods such as feature importance analysis are incorporated to identify the key attributes influencing classification outcomes. Experimental results demonstrate that the proposed framework achieves high detection accuracy while maintaining interpretability, making it suitable for practical cybersecurity applications. By combining strong classification performance with explainability, XAI-Droid contributes to the development of reliable and trustworthy AI-based mobile security systems.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.182

Published by: vikaspatanker

Authors: Mrs. V. Suvarna, Mavuri Bhuvana, Boddu Rajeev, Choppella Vamsi Kumar, Dhulipudi Sree Vivek

Abstract: To improve prediction accuracy and enable real-time monitoring of mechanical components, a deep learning-based approach is proposed. The system utilizes a Convolutional Neural Network (CNN) to extract important features from mechanical parts using sensor data. These features are further processed through fully connected layers for information fusion and classification, allowing accurate prediction of remaining useful life and health status. The trained deep learning model is integrated into a monitoring system to create a complete framework for continuous condition monitoring and life prediction. The system is further optimized to enhance prediction accuracy, real-time performance, and adaptability under different working and environmental conditions. Experimental results show that the proposed model achieves high performance with a Mean Absolute Error (MAE) of 2.1, Root Mean Squared Error (RMSE) of 2.5, and Mean Absolute Percentage Error (MAPE) of 10%. These results demonstrate the effectiveness of the approach and its potential for practical applications in industrial maintenance and reliability improvement.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.183

Published by: vikaspatanker

Authors: Mrs. D. Chakra Satya Tulasi, Bejjipuram Jahnavi, Yarlapati Venkata Naga Durga Varun, Guraja Jayachandra, Peruri Vinay

Abstract: An advanced wild animal detection and alert system is developed using the YOLOv5 (You Only Look Once version 5) model. The system uses an object detection algorithm to identify wild animals and provide real-time alerts to users. A camera captures live video footage, which is processed by a computer running the YOLOv5 model to accurately detect and classify animals. When a wild animal is detected, the system immediately generates alerts such as warning sounds or notifications to prevent potential danger. These alerts can also act as deterrents to scare animals away and improve safety. The system is useful in areas where wild animal movement is common, such as forest borders, agricultural fields, and highways. Overall, the system provides an efficient and real-time solution for monitoring wildlife and reducing human-animal conflicts. Future improvements can focus on increasing accuracy and enhancing real-world performance under different environmental conditions.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.184

Published by: vikaspatanker

Authors: Mrs. A. Daiva Krupa Nirmala, Polina Sai Satvika, Vennapu Lingeswara Rao, Dasari Manvanth, Syed Irfan

Abstract: This paper proposes a smart car parking system that uses image processing and real-time CCTV monitoring to efficiently detect parking space availability and recognize vehicle license plates. The system is designed for both open parking areas and multi-storey parking environments. It uses Python along with the OpenCV library to analyze video input and determine whether parking slots are occupied or vacant based on pixel-level analysis and image processing techniques. In addition, the system integrates Optical Character Recognition (OCR) using the Tesseract engine to automatically extract license plate information from captured images. To improve accuracy, multiple preprocessing techniques are applied to handle variations in image quality, lighting, and noise. The proposed system enables automated parking management, reduces manual effort, and enhances monitoring efficiency, making it suitable for real-time smart parking applications.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.185

Published by: vikaspatanker

Authors: Mrs. K. Harika, Rowthu Kavyanjali Priya, Madeti Vineetha, Samanthakurthy Rajavardhan, Sachin Pandit, Penky Adi Seshu

Abstract: Agriculture plays a vital role in maintaining food security and contributing to the global economy. However, selecting the most appropriate crop for a specific region remains a significant challenge for farmers due to variations in soil nutrients, climatic conditions, and environmental factors. Incorrect crop selection can result in low productivity, inefficient resource utilization, and financial losses. With the growing availability of agricultural data and advancements in artificial intelligence, machine learning techniques have become effective tools for improving decision-making in agriculture. This study proposes an intelligent crop recommendation system that combines machine learning and deep learning models to assist farmers in selecting the most suitable crop based on soil and environmental conditions. The system analyses key agricultural parameters such as nitrogen (N), phosphorus (P), potassium (K), rainfall, soil pH, temperature, and humidity. These features are used to train predictive models capable of recommending the most appropriate crop for cultivation. Various machine learning and deep learning algorithms, including Decision Tree, Random Forest, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Naïve Bayes, Artificial Neural Networks (ANN), Deep Neural Networks (DNN), and Temporal Convolutional Networks (TCN), are implemented and evaluated. The models are trained using a publicly available agricultural dataset containing multiple crop types along with environmental attributes. Performance is assessed using evaluation metrics such as accuracy, precision, recall, and F1-score to identify the most effective model. Experimental results show that ensemble and deep learning models achieve high prediction accuracy in recommending suitable crops. The system also provides a user-friendly interface that enables farmers to input soil and environmental parameters and receive crop recommendations in real time. The proposed approach supports precision agriculture by enabling data-driven farming practices, improving crop yield, and assisting farmers in making informed decisions.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.186

Published by: vikaspatanker

Authors: Mrs. K. Tulya Sree Simla, Penmatsa Dhathri Vidya Prabha, Bikkina Anusha, Gubbala Chandra Mouli, Nimmagadda Sanjith, Vakadi Ayyappa Surya Sri Harsha

Abstract: Traffic violations are a major contributor to road accidents and fatalities, especially in densely populated urban regions. Common violations such as jumping red lights, triple riding on two-wheelers, reckless driving, and riding without helmets significantly increase the likelihood of accidents. Conventional traffic monitoring systems largely depend on manual supervision by traffic police or basic sensor-based methods, which are often inefficient, time-consuming, and susceptible to human error. To overcome these limitations, intelligent traffic monitoring systems based on computer vision and deep learning have gained increasing importance. This paper presents a deep learning-based automated traffic violation detection system using the YOLOv7 object detection model. The proposed system processes video streams captured from roadside surveillance cameras and analyses them frame by frame to detect various traffic violations. The YOLOv7 model is used to identify vehicles and generate bounding boxes around detected objects. A predefined threshold line is applied to determine whether a vehicle crosses the signal during a red light, thereby detecting signal violations. Additionally, the system identifies overloading or triple riding on two-wheelers by analysing the number of riders within a single vehicle bounding box. Helmet violations are also detected by determining whether riders on motorcycles are wearing helmets. If a rider is identified without a helmet, the system classifies it as a violation. The system utilizes publicly available datasets such as the MS COCO dataset for vehicle detection and a custom annotated dataset for detecting overloading and helmet violations. The model is trained and evaluated using performance metrics including precision, recall, F-measure, and mean Average Precision (mAP). Experimental results indicate that the proposed system can accurately detect multiple traffic violations while maintaining efficient real-time performance. The proposed approach offers a cost-effective, automated, and scalable solution for traffic monitoring. It can assist traffic authorities in improving road safety and reducing the burden of manual monitoring. Furthermore, the system can be integrated with existing smart city surveillance infrastructure to support intelligent transportation management and law enforcement.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.187

Published by: vikaspatanker

Authors: Mr. V. Hemanth Sai, Devulapalli Srujan, Mattaparthi Teja Nirgun, Mummidi Lohith Naga Ratan, Poluparthi Abhishek, Kasireddi Naga Venkata Sai Navadeep

Abstract: Social media platforms (SMPs) are widely used for communication and information sharing, but they are also increasingly exploited for criminal activities. These activities include forming illegal groups, spreading false information, stealing personal data, and conducting cyberattacks. The ease of access and anonymity provided by SMPs make them attractive for criminals to perform such actions. Sensitive information such as passwords, financial details, and personal data can be misused, leading to serious threats like identity theft, data breaches, and malware attacks. This paper focuses on detecting criminal activities on social media using machine learning techniques. By analyzing user-generated content, the system can identify suspicious patterns and classify potentially harmful activities. The proposed approach aims to improve early detection and help in preventing cybercrimes effectively. Additionally, it highlights the importance of user awareness and responsible data sharing to reduce risks associated with social media usage.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.188

Published by: vikaspatanker