Authors - Ramesh Chandra Poonia, Vishal Singh Rathore, Ajay Kumar Abstract - In contemporary network infrastructures, ensuring the fidelity of data transmission is paramount for robust communication and security. The intrusion of corrupted data packets can severely degrade network efficiency, resulting in critical data loss, exploitable security gaps, and suboptimal resource allocation. This paper indicates the significantly increase detection accuracy and system resilience by synergistically using the predictive capability of many machine learning paradigms especially. This paper employs sophisticated feature engineering to extract discriminative attributes from network packet headers and payloads, followed by a refined ensemble learning strategy that leverages both stacking and boosting techniques for optimal classification performance. Compared to conventional single-model techniques, evaluated on real-world network traffic datasets our model shows a significant increase in key performance measures. Here a pioneering hybrid machine learning ensemble framework designed for the precise identification and mitigation of corrupted data packets. Notably, the ensemble framework excels in minimizing false positives, enabling real-time packet analysis and bolstering network security. This study contributes to the evolution of intelligent, adaptive network defense mechanisms, providing a scalable and high-performance solution for safeguarding data integrity and mitigating the deleterious effects of corrupted data packets in modern, high-throughput communication environments.
Authors - Smit Patel, Priyanka Patel Abstract - Speech Emotion Recognition (SER) plays a vital role in enhancing human-computer interaction by enabling machines to interpret and respond to human emotions. This study focuses on SER using the RAVDESS dataset, emphasizing speech-only modalities. A comprehensive set of audio features including MFCCs, chroma, spectral contrast, tonnetz, and wavelet transforms is extracted, and the performance of four deep learning models— CNN, LSTM, BiLSTM, and CNN-LSTM—is evaluated. Among them, CNN achieves the highest accuracy (68.37%), with strong F1-scores across several emotion classes. The results underscore the effectiveness of spatial feature extraction in emotion classification and suggest further enhancements using attention mechanisms and transformer-based models.
Authors - Sundar S, Prathilothamai M Abstract - Accurate prediction of monsoon rainfall remains a persistent challenge due to the intricate interplay among meteorological conditions, oceanic influences, and broader climatic patterns. While significant effort has been devoted to improving model architectures, the effect of input feature composition on prediction accuracy has received relatively less attention. This study addresses that gap by conducting an extensive empirical evaluation of 2,047 feature group combi-nations, systematically derived from eleven curated sets of climate-related variables. Using a hyperparameter-tuned XGBoost model, each configuration was evaluated independently to assess the predictive contribution of domains such as lagged climate indices, cyclical temporal encodings, and event-based indicators. The results show that model performance improved significantly from an R² of 0.5801 (RMSE: 10.8999 mm, MAE: 4.9009 mm) using only meteorological features to an R² of 0.7606 (RMSE: 8.2297 mm, MAE: 3.8752 mm) when combined with oceanic and climatic inputs, particularly lagged MJO and ENSO indices and temporal signals. These insights reinforce the value of domain-informed feature fusion and provide a replicable approach to enhancing monsoon prediction models through thoughtful feature group design and empirical validation.
Authors - Atharva Godkhindi, Anjali Naik Abstract - Fraud detection in financial transactions presents a persistent challenge due to extreme class imbalance and evolving attack patterns. While several machine learning (ML) and deep learning (DL) methods have shown promise, these solutions are fragmented and use traditional methods to address the severe class imbalance, leading to models with inflated metrics and poor generalization. In this study, we propose a unified ML-DL-XAI pipeline that integrates Variational Autoencoders (VAE) not only for data augmentation but also for feature engineering. Unlike traditional resampling, the VAE enables representation learning that preserves underlying data distributions while mitigating overfitting. Our pipeline incorporates interpretable machine learning models alongside neural networks to ensure both high performance and explainability. Empirical evaluations on a large-scale financial dataset demonstrate superior and reliable performance, that achieves an accuracy of 99.6%, a precision of 92%, and a recall of 85%, outperforming several recent benchmarks. By combining augmentation, feature engineering, and explainability in a single pipe-line, this work offers a robust and practical answer for real-world fraud detection applications.
Authors - Gitanjali S. Poothuvallil, Dhanya Manayath Abstract - This paper explores the use of generative AI in social entrepreneurship education through the REFLECT model—a justice-informed ethics framework emphasizing participatory, empathic, and power-sensitive reasoning. Using ChatGPT, we simulated context-rich dialogues based on real-life renewable energy cases to examine ethical dilemmas, systemic failures, and power dynamics. Through iterative refinement, we developed a prompt template capable of producing realistic, grounded dialogues. These were evaluated on three criteria: depth of empathy, systems complexity, and presence of ethical tension. Findings suggest that integrating the REFLECT framework enhanced the pedagogical quality of AI-generated cases, aligning them more closely with the aims of values-based social entrepreneurship education. The approach demonstrates how generative AI can foster critical thinking and ethical reflection, helping students engage with complex social justice and sustainability issues. Our study presents a novel application of AI in the classroom, showing its potential to ad-vance pedagogy in socially responsible entrepreneurship. The paper concludes by exploring the future potential of integrating AI into educational pedagogy, highlighting both its ethical implications and the ongoing need for refinement in its application to ensure responsible and effective use in teaching and learning environments.
Authors - Bharateesha lvn, J Vignesh, Jabez Lawrence G, Bhaskarjyoti Das Abstract - The emergence of sophisticated cyber threats calls for the evolution of sophisticated Network Intrusion Detection Systems (NIDS). Even though graph-based approaches have been promising, they have mostly was concerned with node classification to determine the bad actors. This paper provides a new framework that recontextualizes the NIDS challenge as a marginal concern classification problem on a heterogeneous graph. We assume that classifying the boundaries (relations) between network objects as harmful or benign offers a more timely and efficient means of intrusion detection. In order to achieve this, we build a heterogeneous graph from network flow data and employs a Heterogeneous Graph Transformer (HGT) model, which is designed to maintain the integrity of instructional and semantic detail formation present in such graphs. The model is trained and tested on a large dataset from the UNSW-NB15 dataset. Our experiments show that the edge classification method greatly outperforms a conventional node classification baseline, achieving superior accuracy, precision, and recall. These results illustrate the potential of edge-centric GNN models for constructing more efficient and complete network intrusion detection systems.
Authors - Poojitha Panchakarla, Sarvani Kocherlakota Abstract - Cryptocurrency has emerged as a prominent research topic in recent decades, with numerous findings published in leading international journals. To investigate its current research landscape and emerging trends, we conduct a Bibliometric analysis by using R software 4.5.0, and VOSviewer to identify the research trend, emerging topics, and collaborations among countries, authors, documents and the contributing academic journals during the study period. Lecture Notes in Networks and Systems is having the highest publications and IEEE Access is having the most cited document. Kumar A is the most contributing author and Song H is the most cited author. India is the most cited country and contributing country. By research gap analysis, the future direction in the domain of cryptocurrency can be machine language forecasts, sustainable energy, and regulatory framework.
Authors - Jiri David, Jan Fabry, Josef Bradac Abstract - With the growing importance of electromobility, the efficient planning in the production of lithium-ion batteries has become a critical factor in maintaining competitiveness. This article focuses on the optimisation of batch scheduling on parallel processors – an essential challenge in a complex manufacturing environment characterised by hybrid (sequential-parallel) processes, technological dependencies, and high variability. Based on a formal mathematical model of the P|batch, rj , sj | Cmax type, a method is proposed that integrates batch planning, nonlinear setup times, multi-objective optimisation, and robust scenariobased control. The model was implemented in the AMPL (A Mathematical Programming Language) environment and tested using real production data from battery manufacturing at the famous production company. The results demonstrate significant improvements over conventional methods (e.g., FCFS, LPT), including a reduction in production time of up to 10%, a 6% decrease in setup operations, and an increase in capacity utilisation (OEE) by more than 10%. Moreover, reductions in energy consumption and enhanced schedule predictability were observed. The model is designed to be integrable with MES/ERP systems and offers both scalability and adaptability across varying production scenarios. The findings confirm that the combination of a rigorously defined optimisation model and operationally validated data can significantly enhance planning efficiency in battery system manufacturing for electric vehicles.
Authors - Etty Gurendrawati, Hera Khairunnisa, Aji Ahmadi Sasmi, Andrew Saw Tek Wei, Nayla Nandhita Nuril Hadi, Rohadatul Aisy, Shoofiyah Nur Aliifah Abstract - This study examines how Information Quality (IQ), System Quality (SQ), and Service Quality (SeQ) influence Use (U) and User Satisfaction (US). It also investigates the impact of U on US, and how both US and U affect Net Benefit (NB) among Public Sector Accounting students at an Indonesian State University. Data from 114 student questionnaires were analysed using SEM PLS by SMART PLS 4. Seven of nine hypotheses were supported. SQ significantly impacted U and US, while SeQ significantly affected U. Both U and US positively influenced NB, with US having the strongest effect. The insignificant role of IQ suggests students prioritize system functionality and support over data output. These findings emphasize that enhancing system and service quality drives user satisfaction and use, ultimately boosting the perceived net benefits of nonprofit accounting information systems.
Authors - Rhytheema Dulloo, Kirti Biradar, Srijaa M Abstract - The aviation industry faces mounting pressure to achieve net-zero carbon emissions by 2050, yet passenger adoption of sustainable aviation practices remains inconsistent, highlighting the urgent need to understand the psychological and behavioral factors influencing passenger decision-making in sustainable aviation contexts. This study develops and tests an integrated theoretical framework combining Theory of Planned Behavior (TPB), Value-Belief-Norm (VBN) theory, and Technology Acceptance Model (TAM) to explore how environmental values, technology perceptions, and travel context shape attitudes, intentions, and behaviors. A quantitative cross-sectional study of 847 airline passengers across seven metropolitan hubs in India was conducted. Environmental values emerged as the strongest predictor of attitudes toward sustainable aviation behavior (β = 0.42, p < 0.001), while the integrated model explained 62% of variance in behavioral intentions and 34% in actual behavior. Perceived usefulness and ease of use were found to significantly affect sustainable aviation technology acceptance and TPB constructs effectively predicted behavioral intention and actual behavior towards sustainable aviation, with per-sonal norms adding further explanatory power. Significant differences were found between business and leisure travelers, with leisure travelers showing stronger relationships between environmental values and attitudes (p = 0.032) and between intentions and behavior (p = 0.007), while the intention-behavior gap was more pronounced among business travelers, highlighting structural barriers to sustainable aviation adoption. The findings suggest that airlines should adopt differentiated strategies, with sustainability messaging emphasizing personal environmental responsibility for leisure travelers, while structural interventions such as corporate partnerships and policy changes are needed for business travelers to address organizational barriers. This study provides the first comprehensive integration of TPB, VBN, and TAM theories in sustainable aviation contexts, offering novel insights into travel context moderation effects and actionable guidance for industry stakeholders seeking to enhance passenger adoption of sustainable aviation practices.
