Journal of Big Data Analytics and Business Intelligence

An Intelligent Deep Learning and Computer Vision Framework for Automated Camouflaged Wildlife Animal Detection Using YOLOv11 and CAFEM-Lite

Premala Bhande — 2026-06-24

The automated detection of wildlife fauna whose surface patterning and colouration have been refined by millions of years of evolutionary pressure to closely replicate the spectral and textural characteristics of their immediate habitat substrates represents one of the most formidable and consequential unsolved problems at the intersection of computer vision, deep learning, and ecological informatics. Extant general-purpose object detection architectures, engineered and benchmarked predominantly against scenes in which the target object is visually conspicuous against its background, sustain systematic and substantial performance degradation of 20 to 35 % points in mean average precision when confronted with camouflage scenarios, engendering directional bias in population estimates, IUCN status assessments, and anti-poaching surveillance efficacy. This paper presents CAFEM-Lite (Camouflage-Aware Feature Enhancement Module, lightweight variant), a modular, computationally parsimonious image-level preprocessing architecture that amplifies the subtle spectral and boundary signatures by which a camouflaged animal may be discriminated from its background before ingestion by the detection backbone. CAFEM-Lite concatenates three complementary sub-components in sequence—a DCT-based High-Frequency Amplifier (DCT-HFA) that selectively amplifies boundary-encoding high-frequency residuals in the frequency domain; a Boundary-Aware Convolutional Attention (BACA) mechanism that directs channel-level attention toward edge-rich feature regions through Sobel- and Laplacian-derived excitation weights; and a Progressive Context Fusion (PCF) gate that integrates local texture detail with broader semantic context through a learnable channel-partitioned mixing procedure. The entirety of CAFEM-Lite adds only 9,745 trainable parameters. Integrated with YOLOv11n and trained on the COD10K benchmark dataset over 30 epochs on an NVIDIA Tesla T4 GPU, the system achieves validation mAP@0.5 of 0.300, test-set Precision of 65.45%, and inference throughput of 47.3 FPS. A novel Camouflage Difficulty Index (CDI) provides an interpretable per-image quantification of detection hardness, and a browser-accessible Gradio application operationalises the complete pipeline on commodity hardware. The proposed framework constitutes a principled, deployable, and computationally accessible contribution to the nascent but consequential domain of AI-assisted wildlife conservation.

Hierarchical Trust-based Artificial Intelligence Governance Model for Data Protection

Olubodun E. O — 2026-05-28

The widespread adoption of artificial intelligence in organizations has introduced critical risks for trade secret protection, as AI systems may inadvertently disclose or infer sensitive information. Existing governance frameworks lack empirical validation, hierarchical trade-secret safeguards, and integration with zero-trust principles. This study presents a hierarchical trust‑based governance framework that explicitly prevents AI systems from accessing trade secrets beyond authorized organizational levels. A six-equation mathematical model quantifies trust level, access control strength, decision rights, AI usage eligibility, governance effectiveness, and knowledge advancement. The framework is implemented as the HTF Platform, a web application with two-factor authentication, real-time dashboards, and a policy decision confusion matrix. A real-world deployment with 250 audited access decisions achieved 91.6% accuracy, 94.7% precision, 91.6% recall, and 93.1% F1-score. The system architecture and operational flowchart are presented. The confusion matrix confirms that lower-tier users are correctly denied access to higher-tier trade secrets, while legitimate requests are reliably granted. The HTF Platform provides a practical, verifiable, and adaptive solution for safeguarding trade secrets in AI-driven organizations.

An Explainable AI-Driven Healthcare Business Analytics Framework for Intelligent Hospital Management Systems

Md. Royel Islam — 2026-07-03

The rapid digital transformation of healthcare systems, which has generated an unprecedented volume of clinical, operational, and financial data, has simultaneously exposed significant limitations in conventional hospital management systems, particularly in relation to transparency, interpretability, and stakeholder trust in artificial intelligence (AI)-driven decision-making processes. Although AI technologies have demonstrated substantial potential in optimizing healthcare operations, their widespread adoption within hospital management environments continues to be hindered because many predictive models operate as opaque “black-box” systems whose decision-making mechanisms remain insufficiently explainable to administrators, clinicians, and regulatory authorities. To address these challenges, this study proposes a novel Explainable AI-driven Healthcare Business Analytics Framework (XAI-HBAF), which is specifically designed to facilitate intelligent, transparent, and data-driven decision-making within hospital management systems. The proposed framework integrates machine learning algorithms, XAI techniques, and advanced business analytics methodologies in a unified architecture that supports predictive, prescriptive, and descriptive analytics while simultaneously ensuring interpretability and accountability for diverse stakeholders. Furthermore, a hybrid analytical architecture, which combines ensemble learning models with SHAP (Shapley Additive Explanations), has been developed so that both prediction accuracy and model transparency can be significantly enhanced across critical hospital management functions, including patient flow optimization, resource allocation, and financial forecasting. Experimental evaluation, which was conducted using simulated hospital datasets, demonstrates that the proposed framework not only improves prediction accuracy by approximately 18–25% but also reduces operational inefficiencies by nearly 15% when compared with traditional analytics models. In addition, the incorporation of an explainability layer strengthens stakeholder confidence, enhances decision transparency, and supports compliance with evolving healthcare regulations and ethical AI standards. Consequently, the findings of this study suggest that XAI-driven healthcare analytics frameworks can substantially transform conventional hospital management systems into intelligent, adaptive, and trustworthy data-driven ecosystems capable of supporting sustainable and efficient healthcare delivery.

Machine Learning-Based Personalized GATE Performance Prediction System

Chandupatla Varun — 2026-06-22

The Graduate Aptitude Test in Engineering (GATE) is one of the most competitive examinations in India and requires systematic preparation, continuous assessment, and strategic planning. Most existing preparation platforms provide only marks and rankings after mock tests, which often fail to offer meaningful insights into a student’s actual readiness for the examination. To address this limitation, this paper presents a machine learning-based personalized GATE performance prediction system that analyses mock test performance and provides predictive and personalized guidance to aspirants. The proposed system utilizes key performance indicators such as marks, accuracy percentage, subject-wise scores, strong subjects, and weak subjects to estimate future examination outcomes. Machine learning techniques, including regression and Random Forest models, are employed to predict expected GATE score ranges, percentiles, and ranks. The system is developed using a hybrid architecture consisting of a Java Spring Boot backend, a Python-based machine learning module, a MySQL database, and a web-based frontend dashboard. In addition to prediction, the system generates personalized study plans and subject-specific recommendations that help students focus on areas requiring improvement. Experimental evaluation demonstrates that the proposed solution effectively identifies performance patterns and provides actionable insights for better preparation. The system transforms traditional exam preparation into a data-driven, intelligent, and adaptive learning process, thereby improving preparation efficiency and reducing uncertainty among GATE aspirants.

Predictive Analytics in Consumer Buying Behaviour Using Machine Learning

Khushi Bhatia — 2026-05-16

Businesses that want to stay competitive in the digital age need to know how consumers buy things. As more data becomes available from the online world, it is now possible to look more closely at what customers like. Using machine learning and predictive analytics together is a good way to predict what people will do based on past and present data. This study looks at how classification, clustering, and regression machine learning models can be used to predict how people will buy things. It also looks at important factors that affect buying decisions and reviews research that has already been done in this area. The results show that predictive analytics can help businesses better target their customers, personalize their marketing, and make their marketing more effective. Nevertheless, issues like data security, ethics, and model bias should be taken seriously into consideration. Overall, this study has emphasized the significance of predictive analytics as a powerful method to comprehend consumer behaviour and make decisions based on data.