Daily Archives: November 26, 2025

Authors: Srujan Gatla, Teja, Surya, Shashi Kumar

Abstract: Sentiment analysis on social media is a natural language processing technique used to extract subjective information and opinions from user-generated content on various social media platforms, such as Twitter, Facebook, and Instagram. The goal of this project is to perform sentiment analysis on social media data related to a particular topic or brand, such as a product launch or a social issue. Social media data will be collected using relevant APIs or web scraping tools and preprocessed by cleaning and filtering out irrelevant or spammy content. A sentiment analysis model, such as a lexicon-based or machine learning model, will be applied to classify the sentiment of the content as positive, negative, or neutral. Results will be visualized and analysed using various techniques and tools, such as word clouds, bar charts, and time series analysis, to gain insights and make data-driven decisions based on public opinion. Challenges in social media sentiment analysis include the use of slang, emojis, and hashtags, as well as the need to handle multilingual con tent. Sentiment analysis on social media can be a powerful tool for understanding public opinion, customer satisfaction, and brand reputation, and making data-driven deci- sions in various domains, such as marketing, politics, and social sciences.

Published by: vikaspatanker

Authors: Radhika Kulkarni, Tejal Mungase

Abstract: The current job market introduces major difficulties in effectively connecting skilled candidates with suitable employment opportunities. This paper proposes an AI-powered career advisory system using Natural Language Processing (NLP), Machine Learning (ML), and Large Language Models (LLMs) to automate career guidance. The proposed system unites four core modules: semantic-based resume parsing using transformer models, intelligent job matching using BERT embeddings, complete skill gap analysis with customized learning recommendations, and AI-driven resume optimization for Applicant Tracking System (ATS) compatibility. The system implements a three-tier architecture with React.js frontend, FastAPI backend, and a hybrid database layer. This Stage 1 paper presents problem identification, literature survey, system architecture, and methodology. The framework addresses critical gaps in career advisory services through conceptual understanding, customized guidance, intelligent automation, and access to professional career counseling.

Published by: vikaspatanker

Authors: Husna Sultana, Irfan Ahmed, Shivani

Abstract: Interpreting the Urban Black Box, the proliferation of sensors and Internet of Things (IoT) infrastructure in Smart Cities has enabled the development of highly accurate Spatio-Temporal Data Mining models, often relying on deep learning architectures like Graph Neural Networks (GNNs), for tasks such as traffic prediction, crime forecasting, and resource management. Despite their high predictive performance, these models remain "black boxes," hindering their adoption by urban planners and emergency services who require transparency and justification for critical operational decisions. This lack of interpretability poses significant challenges to accountability, auditability, and public trust. This paper addresses the critical need for Explainable AI (XAI) in the urban domain by proposing a novel Spatio-Temporal XAI (ST-XAI) Framework designed for Causal Feature Attribution. Our framework leverages a modified version of SHapley Additive exPlanations (SHAP) combined with the inherent spatial and temporal structure of the data to provide granular, instance-based explanations. The proposed methodology focuses on Temporal Attribution: Quantifying the specific influence of various look-back time windows (e.g., data from the last hour vs. data from 24 hours ago) on the current prediction. Spatial Attribution: Identifying and weighting the contributing influence of specific geographic nodes, links, or neighboring zones within the network structure. Causal Inference: Moving beyond mere correlation by prioritizing features that exhibit a strong, temporally preceding impact, providing a more actionable justification for the prediction. We demonstrate the ST-XAI Framework on a smart traffic prediction model, showing how it successfully translates opaque deep learning outputs into clear, human-understandable narratives. The results illustrate that our framework not only validates model efficacy but also acts as a vital debugging tool for city engineers, transforming black-box predictions into accountable and actionable urban intelligence.

DOI: http://doi.org/10.5281/zenodo.17720787

Published by: vikaspatanker

Authors: Chidanandamurthy G

Abstract: This study compares productivity and carbon performance of organic agroforestry (ORG-AF) and conventional agroforestry (CON-AF) using paired plots under similar soil and climatic conditions. Six pairs of 0.25 ha plots were monitored for three years. System productivity was calculated as the sum of all marketable crop and tree products per hectare, while carbon stocks were derived from tree and crop biomass and soil organic carbon (0-30" " cm). Life cycle inventories of all inputs and field operations were compiled to estimate greenhouse gas emissions and carbon footprints per hectare and per kilogram of product. CON-AF achieved higher system yields (mean 5,808" " kgha^(-1)) than ORG-AF (mean 5,017 kgha^(-1)), a difference of about 16%. In contrast, tree biomass increment was greater in organic plots (3.55tha^(-1) yr^(-1)) than in conventional plots ( 2.55tha^(-1) yr^(-1)), and soil carbon increased faster in ORG-AF (0.43tCha-1yr^(-1)) than in CON-AF (0.16tCha^(-1) yr^(-1)). Total annual carbon stock change averaged 2.09tCha^(-1) yr^(-1) in ORG-AF and 1.36tCha^(-1) yr^(-1) in CON-AF. Area-based carbon footprints were 2,950 and 4,150" " kgCO_2-eq ha^(-1) yr^(-1) for organic and conventional systems, respectively, while product-based footprints were 0.59 and 0.71" " kgCO_2-eq kg^(-1). Both systems acted as net carbon sinks, but net carbon balance was much higher in ORG-AF (4.7vs.0.8tCO_2-eq ha^(-1) yr^(-1)). The results show that organic agroforestry can maintain high productivity while substantially improving carbon efficiency and climate mitigation potential.

DOI: http://doi.org/10.5281/zenodo.17720645

Published by: vikaspatanker

Authors: Dr. Latha H R

Abstract: Secure communication and computing is the vital requirement of the day as global networks and information systems are expanding like the big bang theory of the universe. People have started treating information as an asset. The information asset needs to be secured from attacks. Everything in the world is being upgraded to electronic communication and this requires protection against data fraud. Information has chosen different media like text, image, audio, video and multimedia for its existence. Cryptography is the science which provides techniques for securing information over network. Network security is the process of taking physical and software measures to protect underlying infrastructure. This paper introduces cryptography, chaotic cryptography, its computational power in image security. It proposes new sealion algorithm to increase the computational power of cryptographic algorithms. It also verifies the efficiency of proposed algorithm against benchmarks set for the security of images over network.

DOI: http://doi.org/10.5281/zenodo.17720340

Published by: vikaspatanker

Authors: Indu Shinde, Sarvada Anvekar, Purva Nargide, Shravni Shikhre, Shantu Pujari, Professor P.S.Pandhre

Abstract: Access to legal information and assistance remains a major challenge for many individuals due to high costs, lack of awareness, and geographical barriers. With the rapid growth of Artificial Intelligence (AI) and Natural Language Processing (NLP), chatbots have become an effective tool for improving access to information and services. This research paper presents the design, development, and evaluation of a Legal Aid Chatbot that provides preliminary legal guidance to users in a simple and accessible manner. The system uses NLP techniques, machine learning models, and a structured legal knowledge base to understand user queries and generate meaningful responses.

Published by: vikaspatanker