Category Archives: Uncategorized

Authors: Jennifer Roberts, Rebecca Turner, Victoria Hughes, Richard Morgan, Chaitanya Srinivas, Akhilesh Achari

Abstract: Cloud-native architectures have become the foundation of modern digital applications due to their scalability, flexibility, resilience, and ability to support continuous deployment across distributed computing environments. However, the increasing complexity of microservices, containers, orchestration platforms, and dynamic workloads introduces significant challenges in maintaining system reliability and preventing service disruptions. Traditional reactive maintenance approaches often fail to identify potential failures before they impact application performance and user experience. This research presents a predictive failure analysis framework for reliability engineering in cloud-native architectures that leverages predictive analytics, machine learning algorithms, real-time monitoring, and intelligent fault detection mechanisms to proactively identify and mitigate system failures. The proposed approach continuously analyzes operational metrics, infrastructure logs, service dependencies, and workload patterns to detect anomalies, forecast potential failures, and recommend corrective actions before critical incidents occur. By integrating predictive models with cloud-native reliability engineering practices, the framework supports automated fault diagnosis, resource optimization, resilience enhancement, and service continuity. The study explores key architectural components, predictive techniques, reliability metrics, and implementation strategies for building highly available and fault-tolerant cloud-native systems. Experimental evaluation demonstrates improvements in failure prediction accuracy, system uptime, response performance, and operational efficiency compared to conventional monitoring methods. The findings indicate that predictive failure analysis provides a robust foundation for developing intelligent, adaptive, and resilient cloud-native infrastructures capable of supporting the growing demands of modern enterprise applications and distributed computing ecosystems.

DOI: http://doi.org/10.5281/zenodo.20797204

Published by: vikaspatanker

Authors: Jidnyasa Bagul, Pradnya Dhavan, Project Guide Dr. Nandini Dhole

Abstract: In modern institutions and organizations, effective communication of information is essential, yet traditional notice boards often fail to provide timely updates and accessibility. This project presents a Voice-Based Notice Board system that leverages speech recognition and text-to-speech technologies to automate the process of publishing and delivering notices. The system allows authorized users to input notices through voice commands, which are then converted into text using speech-to-text processing. The processed information is stored in a cloud-based database and can be displayed on a digital screen as well as broadcasted through audio output using text-to-speech synthesis. The integration of Internet of Things (IoT) technology ensures real-time updates and remote accessibility. The proposed system is built using Raspberry Pi, along with a microphone module, speaker system, and display unit. Python is used as the primary programming language, incorporating various speech processing libraries for accurate voice recognition and natural audio output. This solution enhances accessibility, reduces manual effort, and ensures faster dissemination of information. It is particularly useful in environments such as educational institutions, offices, hospitals, and public spaces, where quick and efficient communication is crucial. The system also provides scope for future enhancements, including multilingual support and mobile application integration. Overall, the Voice-Based Notice Board offers a smart, efficient, and user-friendly alternative to traditional notice systems by combining automation, cloud computing, and voice interaction technologies.

Published by: vikaspatanker

Authors: Bharti Saxena, Rupali Chaure, Ashish Chourey, Mohit Singh Tomar

Abstract: Here we introduce an empirical exploration of a real-time Hybrid Deep Learning model for Age and Gender Recognition (HDL-AGR) based on facial images collected from multiple unconstrained scenarios. Estimate age and gender from facial images is a classic computer vision problem with applications ranging from human-computer interaction, intelligent surveillance, personalized marketing to healthcare screening. Most existing approaches are limited by low accuracy on far-side age groups, extreme sensitivity to lighting and occlusion, and extreme computational overhead that would preclude real-time deployment. The proposed HDL-AGR framework consists of a backbone (which has been defined as a modified EfficientNet-B4 convolutional base), attention module (Transformer-based), and an output head (dual-branch, trained jointly for age regression and gender classification) to be tuned up to date. The model is trained and evaluated with five benchmark datasets UTKFace, IMDB-WIKI, Adience, CACD and Fair Face containing over the 845K annotated images. Empirical results: HDL-AGR achieves. (i) A new state-of-the-art Mean Absolute Error (MAE) of 3.94 years in age estimation, along with an unprecedented gender classification accuracy of 97.2% and (ii) Operates at an inference speed of 54 frames per second on standard GPU hardware – outperforming all compared peer methods in the process. The contribution of each architectural component is confirmed through ablation studies. Conclusion: Our results identify HDL-AGR as a strong, efficient, and practically deployable approach for online recognition of facial attributes.

DOI: http://doi.org/10.5281/zenodo.20792037

Published by: vikaspatanker

Authors: Radhika Kulkarni, Tejal Mungase, Prof. Shradha Pawar

Abstract: Choosing the right career path and the right job opportunity has become increasingly difficult in a labour market where industry requirements evolve faster than academic curricula and where the sheer volume of job postings makes manual evaluation impractical for most candidates. This paper presents the AI-Based Career Advisor, an intelligent system designed to help individuals understand how well their resume aligns with a target job description, identify missing skills, and receive concrete, personalized guidance for improving their employability. The system combines a supervised machine learning model with natural language processing and large language model components to deliver this guidance in a single, integrated workflow. At its core is a resume–job description fit classifier trained on 6,241 real-world resume–job pairs sourced from a public dataset, using TF-IDF based feature engineering across 10,012 dimensions. Six candidate algorithms — Logistic Regression, Naive Bayes, Support Vector Machine, Random Forest, a Neural Network, and XGBoost — were trained and compared, with XGBoost emerging as the best-performing model after hyperparameter optimization, achieving 78.14% test accuracy and an 89.57% ROC-AUC score. The system further incorporates a hybrid skill-extraction pipeline built on spaCy's named entity recognition and phrase matching, a GPT-4-based resume enhancement module accessed through LangChain, and supporting modules for learning-resource and project-idea recommendation. The complete pipeline is deployed as an interactive Streamlit web application, giving users real-time predictions and actionable career feedback. This paper discusses the motivation, design, methodology, and evaluation of the system, and outlines directions for extending it into a more comprehensive career guidance platform.

Published by: vikaspatanker

Authors: Mrs. Bhumika Alte, Satyam Mali, Yashraj Mhase, Kishor Hirgal

Abstract: Aspect-Based Sentiment Analysis (ABSA) provides a fine-grained approach to understanding opin-ions by extracting aspect–opinion–sentiment relation-ships from text. It is particularly valuable in domains such as product reviews, customer services, banking, and social media, where identifying specific strengths and weaknesses is essential. The subtask of Aspect-based Sentiment Triplet Extraction (ASTE) extends ABSA by simultaneously identifying aspect terms, corresponding opinion expressions, and their sentiment polarities. This work proposes an improved ABSA framework that integrates pre-trained language models (PLMs) with a pruned syntactic encoding mechanism to efficiently capture both local and global contextual dependencies. Additionally, a dynamic encoding strategy is introduced to overcome the limitations of traditional local encod-ing, which often fails to capture long-range relation-ships between aspects and opinions. The combination of syntactic pruning and dynamic encoding enhances the association between aspect and opinion terms, leading to more accurate sentiment classification. Experimental evaluations on benchmark ABSA datasets are expected to demonstrate that the pro-posed model achieves higher accuracy and robustness compared to existing methods. This approach effec-tively combines syntactic structure and contextual un-derstanding, improving interpretability and performance in aspect-level sentiment prediction tasks.

DOI: http://doi.org/10.5281/zenodo.20787176

Published by: vikaspatanker

Authors: Hitesh Jitendra Jadhav, Santosh Shriram Karvar, Atharv Arun Patil, Gaurav Anil Waje, Gaurav Vijay Barde, Bajirao Subhash Shirole

Abstract: A significant percentage of traffic accidents in the world result from sleepy drivers. Although a number of detection methods have been established, their utility is often problematic. Physiological signals (EEG, ECG) and vision- based behavioral cues (eye closure, yawning) have been studied in the past, and deep learning models such as Convolutional Neural Networks (CNNs) have shown excellent accuracy in controlled settings. Significant gaps still exist, though, especially in the areas of robustness against various lighting conditions and occlusions, validation in on-road scenarios, and non-intrusive, computationally efficient systems appropriate for real-time deployment on mobile platforms. This review highlights the shortcomings of current vision-based approaches while synthesizing and critiquing them. It then suggests a future- focused approach based on a lightweight CNN architecture (like MobileNetV2) optimized for on-device inference with TensorFlow Lite. This work attempts to close the gap between academic research and useful, scalable solutions that can improve road safety by concentrating on a camera-based, non – intrusive system deployable on common Android devices.

DOI: http://doi.org/10.5281/zenodo.20786979

Published by: vikaspatanker

Authors: Rasika Kokate, Saloni Gohad, Vaishnavi Gulave, Tanuja Karpe, Sunita Borse

Abstract: AYUSH (Ayurveda Yoga Naturopathy Unani Siddha and Homeopathy) system is a repository of the wisdom obtained from 8000 plants. But most of this knowledge is available in printed and handwritten Sanskrit and Hindi manuscripts which are computing unfriendly. This study introduces an end-to-end AYUSH knowledge recommendation pipeline based on AI to digitize, interpret and recommend insights from the AYUSH body of knowledge for modern computational intelligence. The framework combines Optical Character Recognition (Tesseract OCR), NLP for Indic languages, Knowledge Graph modelling (Neo4j) and AI-based reasoning (BERT, Random Forest) to convert unstructured manuscripts into searchable knowledge that can be analyzed by human . The system captures herbal, disease and treatment entities, relates the entities semantically, and then provides query-driven recommendations through an intelligent interface. Using a simple interface, researchers would be able to ask for insights such as “What are the herbs that have been associated with anti-inflammatory activity?” This strategy lowers the expense of early stage drug discovery, validates traditional remedies, and forges new roads in integrated health care investigation. This study provides the infrastructure for AI- based analysis of literature on traditional medicine and adds to digital conservation, availability and edification as well as evidence-informed integrated healthcare.

DOI: http://doi.org/10.5281/zenodo.20786139

Published by: vikaspatanker

Authors: Bhuvanesh G, Arun Das, Shivanand Chinnappanavar, Ravikumar M

Abstract: This study aimed to assess suitable lands for agricultural purposes in the Gundlupet taluk of Chamarajanagar district. Leveraging the widely used Analytic Hierarchy Process (AHP) integrated with Geographic Information System (GIS), this research conducted a thorough land use suitability analysis. Key parameters including geomorphological and geological features, relief, slope, drainage density, rainfall, soil texture, and land use and land cover were considered in the analysis. Weights were assigned to these parameters based on their significance and importance, resulting in the generation of an agricultural land suitability map divided into three categories. Upon excluding forested and reservoir areas from the reclassified suitability map, the study estimated that 19.59% of the study area (266 sq. km) is highly suitable for agricultural production, 67.6% (918 sq. km) is moderately suitable, and 12.81% (174 sq. km) is unsuitable for agricultural production in this region. This framework facilitates the early zoning of agricultural land for protection, ensuring sustainable land use development in the future.

DOI: http://doi.org/10.5281/zenodo.20783644

Published by: vikaspatanker

Authors: Lohitha Lakshmi K, Hema Sri S, Shaik Reshma, Hima Sai Nandhan P, Manoj Kumar Reddy S D V

Abstract: Data plays a key role in analysis and machine learning, but working with real-world datasets is often challenging because they usually contain missing values, duplicate entries, inconsistencies, and noise that can affect the accuracy of results. Data cleaning is therefore an essential step, yet it can be time-consuming and often requires programming knowledge, making it less convenient for many users. In this work, we present DataForge, a data preprocessing system designed to make the cleaning process simpler and more accessible. The platform allows users to upload datasets and perform cleaning operations without writing code, using a mix of statistical methods and simple intelligent techniques to handle issues such as missing data, outliers, and duplicate records. Overall, DataForge focuses on reducing the effort required for data preparation while still helping users work with more reliable datasets. This approach also helps users get a clearer idea of their data without going into too much technical detail.

Published by: vikaspatanker

Authors: Ms. Snehal Kadbhane, Mr. Ritesh Khandagale, Dr. Vijaykumar Kale, Dr. Mahesh Thakare, Vaibhav Narwade

Abstract: Background: The near-universal reliance on high-sucrose vehicles in paracetamol oral syrups creates an increasingly untenable clinical tension for vulnerable patient populations—diabetic individuals experiencing glycemic excursions, children at heightened risk of dental caries, and obese or metabolically compromised patients. With global diabetes prevalence now exceeding 537 million adults and dental caries ranking as the world's most prevalent non-communicable condition, the pharmacoeconomic and public health argument for sugar-free alternatives has become irrefutable. Methods: Five trial formulations (F1–F5) of a sugar-free paracetamol oral syrup at 120 mg/5 mL were developed using a Quality by Design (QbD) framework. Sorbitol (20–30% w/v), hydroxypropyl methylcellulose K4M (0.25–0.75% w/v), and sucralose (30–70 mg/100 mL) were systematically varied while all other excipients were held constant. Formulations were evaluated for organoleptic acceptability, pH, viscosity, drug content, density, surface tension, sedimentation ratio, and antimicrobial preservative effectiveness per USP <51> Category 2. The optimized formulation (F3) underwent 90-day accelerated stability testing per ICH Q1A(R2) at 40°C ± 2°C/75% ± 5% RH and was benchmarked against a commercially marketed sugar-free reference product. Results: F3, containing sorbitol 25% w/v, HPMC K4M 0.50% w/v, and sucralose 50 mg/100 mL, emerged as the optimized formulation. It exhibited a pH of 5.82 ± 0.02, viscosity of 92 ± 2.5 cps, drug content of 99.4 ± 0.5% of label claim, and a palatability score of 4.5/5.0—superior to both lower-concentration variants and the marketed comparator (4.2/5.0). Accelerated stability studies confirmed drug content above 98.6% and p-aminophenol below 0.08% at day 90, well within pharmacopoeial limits. All five challenge organisms met USP <51> Category 2 acceptance criteria. Conclusion: The optimized sugar-free paracetamol syrup demonstrates pharmacopoeial compliance, chemical and microbiological stability supportive of a 24-month shelf life, and patient acceptability equivalent or superior to a marketed reference. The formulation strategy—combining a polyol bulk sweetener with a high-intensity non-caloric sweetener and a cellulose-ether viscosity modifier—provides a scientifically validated, clinically advantageous platform for analgesic-antipyretic therapy in patient populations for whom conventional sucrose-based preparations are contraindicated or undesirable.

Published by: vikaspatanker