International Journal of Data Science and Advanced Analytics

Insect Pest Classification Using Vision Transformers

Remen Khullar — 2025-10-23

Transformer-based neural networks have emerged as the dominant approach for natural language processing tasks due to their strong performance in text understanding and generation. Motivated by this success, recent research has explored the application of transformer architectures to computer vision, an area traditionally dominated by Convolutional Neural Networks (CNNs). Although the adoption of transformers in vision tasks is relatively recent, early studies have demonstrated promising results, particularly in image classification. However, limited research has systematically evaluated Vision Transformer models for fine-grained insect pest classification using large-scale agricultural datasets. In this paper, a Vision Transformer (ViT)-based computer vision model is proposed for the automated identification and classification of insect pests that damage agricultural crops. Insect pests destroy nearly one-third of global agricultural production and pose significant risks to food security and public health. Automating pest identification using deep learning can significantly reduce the time and effort required compared to manual inspection. The proposed model is trained and fine-tuned on the IP102 dataset, which contains more than 75,000 images spanning 102 categories of common crop-damaging insect pests. Experimental results demonstrate that the 8×8 ViT achieved 41.75% test accuracy, 67.28% top 5 accuracy, and 85.56% ROC-AUC, compared with 47.61% accuracy, 73.76% top 5 accuracy, and 91.21% ROC-AUC achieved by the best-performing CNN, InceptionV3. These results demonstrate the potential of transformer-based architectures for fine-grained insect pest classification in the IP102 benchmark.

Leveraging Advanced Machine Learning Algorithms for Optimized Supplier Selection: A Multi-Criteria Decision-Making Approach

NIKHIL NAN — 2026-06-09

This study investigates the application of K-means clustering, a machine learning technique, to enhance supplier selection and management within large organizations. By segmenting suppliers based on key performance metrics, the research aims to establish more strategic supplier relationships. The methodology leverages purchase order data transformed for clustering analysis. K-means clustering is chosen for its effectiveness in grouping suppliers with similar characteristics. The analysis identifies four distinct supplier segments: high-ranked, steady growth, high spending, and rapid growth. These segments become the foundation for developing targeted supplier management strategies. The study explores practical implications like retention programs for high-ranked suppliers, growth support for steady performers, cost management for high spenders, and investment opportunities for rapidly growing suppliers. Additionally, it discusses supplier rationalization and the advantages of data-driven decision-making in supplier relationship management. The concluding remarks emphasize the effectiveness of K-means clustering for supplier segmentation. The research offers a framework for optimizing supplier selection and management, with the potential to support improved efficiency, stronger supplier relationships, and better business outcomes through more targeted and data-driven supplier management.

Uncertainty, Disagreement, and Information Fusion in Modern AI Systems: A Comprehensive Survey

Mohammed Kasra Sartaee — 2025-10-10

The reliability of modern artificial intelligence systems critically depends on their ability to quantify uncertainty, interpret disagreement, and fuse information from multiple models or data sources. While classical machine learning emphasized probability calibration for discriminative classifiers [10], contemporary AI systems increasingly operate in open-ended, generative, and multimodal settingswhere uncertainty is semantic, subjective, and decision dependent [17]. This survey provides a comprehensive review of uncertainty estimation, disagreement modeling, and information fusion in modern AI systems. We formalize aleatoric and epistemic uncertainty [4] and review foundational evaluation metrics including negative log-likelihood, proper scoring rules, and calibration error [8]. We then survey uncertainty estimation methods spanning Bayesian neural networks [5], [22], deep ensembles [18], calibration techniques [10], consistency-based approaches [27], and semantic uncertainty measures for generative models [17]. Disagreement is analyzed as an informative signal rather than noise, covering ensemble disagreement [12], human annotation variability, and multi-agent systems. Finally, we review uncertainty-aware information fusion techniques for multimodal and multi-model systems [6]. Throughout the survey, we highlight empirical trade-offs between computational cost and reliability, discuss evaluation practices, and identify open challenges related to scalability, distribution shift, hallucination detection, and decision-theoretic integration. This work aims to serve as a unified reference for researchers and practitioners designing AI systems that are not only accurate, but reliably aware of their own limitations.

Digital Transformation as a Catalyst for Activity-Based Funding Implementation

Ahmed Emad Alhamalawy — 2025-09-01

Activity-Based Funding (ABF) has gained traction globally to enhance hospital reimbursement by linking payments to volume and complexity of care via Diagnosis-Related Groups (DRGs). Empirical evidence on ABF’s effectiveness remains mixed and context dependent. Notably, the Middle East and North Africa (MENA) region lacks rigorous analyses of ABF implementation due to lack of understanding of advanced digital health technologies. This study investigated the role of digital transformation in ABF implementation within Saudi Arabia’s healthcare system considering institutional, organizational, and technological factors. A comparative qualitative case study was employed, analyzing policy documents, implementation reports, and stakeholder interviews. The study applied the framework of the principal-agent theory to interpret incentive structures and utilized a digital maturity lens to assess the integration of health information systems supporting ABF. Findings revealed that ABF reforms were initiated and aligned with national development agendas. Saudi Arabia’s relatively advanced digital health ecosystem has facilitated more effective cost accounting, data integration, and transparency, enabling smoother ABF rollout. This Saudi context underscores the necessity of aligning stakeholder incentives and building organizational capabilities for data-driven decision-making to realize ABF’s efficiency and quality improvement goals. In conclusion, this study contributed to the research dialogue by demonstrating that successful ABF implementation in emerging economies surpasses financial restructuring per se, to leverage digital transformation to overcome systemic barriers. Policymakers are advised to prioritize investments in digital health infrastructure and governance frameworks to support sustainable, value-based healthcare financing reforms in the MENA region.

AI-Driven News Summarisation for Financial Insights: Revolutionising Large Language Models

Shraddha Bhoir — 2025-09-19

Financial professionals rely on timely news to support decision making. However, manually reviewing large volumes of financial news is inefficient particularly when articles are complex and lengthy. Automated text summarisation using large language models (LLMs) offers a promising solution for summarising extensive textual information. This study aims to customise and optimise a text summarisation model for financial news. To achieve this, the study examines the effectiveness of transformer based LLMs by fine-tuning the FLAN-T5-XL model for this domain. Experiments were conducted using a dataset of 2,000 general news articles. Performance was assessed using ROUGE metrics and expert human evaluation. The results show that the fine-tuned FLAN-T5-XL with truncation achieved the best performance obtaining a ROUGE -1 score of 55 and 86% agreement with expert evaluation. These findings demonstrate that domain adapted LLMs can provide a practical tool for rapid information synthesis and financial decision making.