The evolution of audit technology is changing the profession, but studies conducted in Kuwait and the GCC have remained ad hoc, tending to treat traditional audit software and AI analytics as two separate entities rather than as a unified technological trajectory. This research addresses that gap by analysing the interaction of these tools as a sequential process of improving audit quality. A mixed-methods design was employed, combining a structured survey of 219 auditors and finance executives with semi-structured interviews. Analysis proceeded through hierarchical descriptive tabulation, factor validation testing, structural equation modelling (PLS-SEM), bootstrapped mediation and moderation testing, incremental value analysis, and multi-group comparison. The findings demonstrate that audit software and AI analytics are complementary rather than competing technologies: software improves process efficiency and compliance foundations, while AI analytics extends these foundations through predictive risk capabilities and fraud detection maturity. Auditor expertise, targeted training, and organisational readiness significantly moderate both pathways. Adoption of both tools in combination produced the strongest gains in audit quality, and multi-group analysis revealed contextual differences across GCC firms. This paper makes three contributions. First, it provides empirical validation of a Sequential Adoption Model, demonstrating that audit software and AI analytics are complementary and sequentially ordered phases of a single audit technology trajectory. Second, it identifies auditor expertise, targeted training, and organisational readiness as key moderators of both pathways, and documents significant contextual differences between Kuwaiti and broader GCC firms. Third, it establishes a planning-phase boundary condition: AQF-based evidence from the GCC indicates that the planning dimension (AQF 2) remains the least effectively technology-supported phase even after sequential adoption, pointing to a phase-specific gap whose mechanisms are examined in complementary conceptual work.

Accurate Remaining Useful Life (RUL) prediction is essential for reliable battery management and cost-effective maintenance in lithium-ion energy storage applications. This study proposes a leakage-safe learning framework for battery RUL prediction from numerical cycling features, emphasizing realistic generalization to unseen batteries through group-aware splitting. The core contribution is a proper out-of-fold (OOF) stacking ensemble trained under GroupKFold to prevent information leakage across samples from the same battery while enabling an effective meta-learner to combine diverse learners. Experiments are conducted on a public Battery RUL dataset from Kaggle, and performance is evaluated using complementary regression metrics (R², MAE, RMSE) and robust percentage errors (sMAPE, WAPE). Results show that the proposed stacking approach (OOF GroupKFold with Meta-XGBoost) achieves the best overall performance (R² = 0.999550, MAE = 5.297907, RMSE = 6.834146, sMAPE = 3.080569, WAPE = 0.958618), outperforming strong baselines including XGBoost and Random Forest. These findings confirm that leakage-safe group-aware stacking can significantly enhance accuracy and stability for battery RUL prediction in practical deployment settings.

BC innovation is increasingly recognized in the accounting domain as a transformative technology with significant implications for accounting practices. This research aims to examine the potential digital transformations in Saudi businesses regarding the integration and adoption of BC technology into accounting practices. To analyze the associations between variables, the research design employs a quantitative method using PLS-SEM by Smart-PLS. Data were collected from a convenience sample of 208 respondents employing a standardized questionnaire. The results reveal that budgeting processes, transaction recording, and accounting adjustments, as well as BC's awareness, are positively influenced by the adoption of BC innovation. The findings, however, indicate that accounting operations should further support their innovative capabilities to remain aligned with rapid technological advancements. These results contribute to the existence of a stock of knowledge on the use of BC in accounting and auditing, also delivering empirical evidence of its benefits in an emerging nation The study has practical implications for accountants, businesses, policymakers, and researchers. The results can help businesses in efficiently gaining an advantage from BC developments to promote their accounting processes. Policymakers can adopt these outcomes to establish supporting regulations and frameworks that motivate the adoption and integration of BC innovation in the domain.

This study develops an explainable ensemble machine learning framework for sustainable credit risk assessment in peer-to-business (P2B) lending, a rapidly expanding FinTech model that enhances access to financing for small and medium-sized enterprises (SMEs). The increasing reliance on algorithmic decision-making underscores the need for transparent and interpretable credit evaluation, particularly in environments characterized by information asymmetry and reliance on borrower-provided disclosures. To address these challenges, a heterogeneous ensemble model is proposed, integrating Random Forest (RF), Light Gradient Boosting Machine (LGBM), and deep learning classifiers within a soft-voting architecture. Feature selection and class balancing are guided by RF importance scores and resampling techniques, resulting in a compact and interpretable 12-feature set comprising pricing, contractual, and transaction-level variables derived from borrower disclosures. Using real-world transaction-level data from a UK platform, the proposed model achieves improved predictive performance (ROC-AUC = 0.767) compared to a neural network baseline (ROC-AUC = 0.717) under severe class imbalance. SHAP-based explainability analysis identifies Maturity Days, Annualised Gross Yield, Advance Rate, and Discount Rate as the most influential predictors, highlighting how disclosed information is translated into pricing and contractual terms in digital lending markets. The findings demonstrate that disclosure-informed features can enhance both predictive accuracy and interpretability, supporting more transparent, robust, and accountable credit risk assessment in FinTech lending environments.

Recent developments in single-cell and high-throughput sequencing technologies have enabled the collection of genomic data at multiple time points; therefore, the investigation of dynamic processes and the observation of disease prognosis, progression, and therapy effects are now possible. Yet, challenges, including biological variability, high dimensionality, noise, and temporal resolution in these datasets, need to be addressed. This review aims to categorize the machine learning (ML) methods used to create models that analyse the temporal dynamics of genomic data. In addition, the best way to apply either the classical or modern deep learning approaches to analyse temporal genomic and multi-omics datasets is being discussed. ML methods used in time-series genomics analysis have been organized by data type and modeling approach to compare their efficiency in handling sparsity and nonlinear temporal relationships, thereby helping to understand the biological mechanisms underlying them. Furthermore, we focus on the recent frameworks that combine temporal learning models with long-read and single-cell sequencing. While ML provides robust tools to reveal temporal dynamics, data standardization, scalability, interpretability, and benchmarking remain challenges. We summarize best practices for model evaluation and outline future directions, including multi-omics integration, interpretable artificial intelligence, and large-scale, reproducible benchmarks. Via combining computational ingenuity and biological insight, we enable a deep understanding of the complexity of biological processes and pave the way for applying precision medicine through ML-based time-series genomics.

Spectral methods are among the most powerful approaches to recovering global structure in network data, but the vast majority of existing theory is developed for crisp graphs, in which vertices and ties observed are unambiguous. Both the vertices and edges are uncertain in terms of existence, strength, and reliability in many network-science contexts, such as social-media interaction data (friendship networks etc), trust networks, recommendation systems and partially observed relational databases. This research introduces a mathematically principled spectral theory for undirected fuzzy graphs, which is based on vertex-normalized fuzzy adjacency matrix and the corresponding Laplacian operators. The construction has symmetry, obeys the fuzzy constraint μ_ij ≤ min(σ_i, σ_j) and when all vertices belong to one membership class reduces down to classical weighted-graph adjacency and Laplacian. Under this paradigm several structural outcomes are proven: the fuzzy Laplacian is positive semidefinite, its nullity matches the number of connected components of the support graph, and the second Laplacian eigenvalue measures fuzzy algebraic connectivity, along with explicit upper and lower spectral bounds. A cut-based inequality and perturbation theorem are subsequently obtained to characterize community separability, as well as robustness against membership noise. The theory is exemplified on a six-node fuzzy network, and then the new method is applied to the well-known Zachary karate club benchmark after equitable fuzzification of vertices and edges. In the empirical study, we present that with this method the canonical split is recovered with 94.12 % accuracy, a fuzzy modularity of 0.3645 is produced and bridge-like actors are identified based on Perron fuzzy centrality score along with very high stability under multiplicative perturbations of edge memberships. The paper thus provides a rigorous spectral toolkit for uncertainty-aware network analysis, filling an important bridge between fuzzy mathematics and modern data and network science.

This study examines the influence of AI-driven personalization capabilities on consumer purchase intention, focusing on the mediating role of customer engagement. It explores how recommendation accuracy, real-time adaptation, and predictive analytics capability enhance customer engagement in digital commerce. Data were collected from 473 online consumers using a questionnaire and analyzed using Structural Equation Modeling with SmartPLS. The findings reveal that all three AI-driven personalization dimensions significantly and positively affect customer engagement. In turn, customer engagement has a strong positive impact on purchase intention. Moreover, customer engagement was found to mediate the relationships between AI-driven personalization capabilities and purchase intention, confirming its critical role in translating technological capabilities into consumer behavior. All hypotheses were statistically supported. This study enriches the literature on AI in marketing context by clarifying how personalization capabilities influence consumer decisions. It also provides practical insights for businesses aiming to enhance engagement and increase purchase intention through advanced AI-driven personalization strategies.

The increasing complexity and frequency of cyberattacks have made accurate and reliable intrusion detection systems (IDSs) essential for modern network security. In this study, an explainable triple-hybrid deep learning framework is proposed for binary intrusion detection using the CICIDS2017 dataset. The proposed architecture integrates three complementary branches, namely a Transformer encoder, a bidirectional long short-term memory (BiLSTM) network, and a multilayer perceptron (MLP), to capture global feature interactions, sequential dependencies, and nonlinear discriminative patterns from network traffic data. To enhance adaptive representation learning, the framework employs a branch-gating mechanism and a fusion-gating module before final classification. The model was evaluated in a Benign-versus-Attack setting using 5-fold stratified cross-validation and assessed through accuracy, precision, recall, F1-score, ROC-AUC, confusion matrices, and explainability analysis. Experimental results showed strong and stable performance across folds, with a mean validation accuracy of 97.18%, a best-fold accuracy of 97.43%, and a mean ROC-AUC of 0.9975. LIME-based explanations further improved transparency, confirming the framework as an effective and interpretable solution for binary intrusion detection.

This study investigates the determinants of online learning application effectiveness in Vietnam using a structural equation modeling (SEM) approach. Grounded in the Unified Theory of Acceptance and Use of Technology (UTAUT) and extended with psychological constructs, the proposed model integrates performance expectancy, effort expectancy, social influence, facilitating conditions, learning habits, intrinsic motivation, technostress, and feeling of isolation. Data were collected from 501 online learners and analyzed using SPSS and AMOS. The results reveal that performance expectancy, effort expectancy, learning habits, intrinsic motivation, and technostress significantly influence behavioral intention, while social influence and facilitating conditions are not significant. Behavioral intention strongly predicts actual usage behavior. Notably, technostress emerges as the most influential factor, suggesting that technological pressure may act as both a barrier and a driver of engagement depending on user adaptation. The study contributes to the literature by integrating enabling and inhibiting factors into a unified framework and offers practical implications for improving online learning systems in developing countries.

With the constantly growing interaction of artificial intelligence and green financial technology, the emerging possibilities introduced new possibilities to optimize organizational performance in the energy sector as business entities were being pressured to be more financially efficient and strive towards environmental sustainability. Although earlier research explored the role of digital technologies in financial services and sustainable business practices, limited empirical attention was given to how artificial intelligence adoption affects Green FinTech performance and whether sustainable innovation mediates this relationship. The paper has explored the effects of the introduction of artificial intelligence on Green FinTech performance in the energy industry and explored the mediating effect of sustainable innovation. The research design used was quantitative and data was gathered in firms working in the energy industry. The relationships proposed were analyzed with the help of suitable statistical tools to evaluate the direct impact of artificial intelligence implementation on Green FinTech performance and the indirect impact via sustainable innovation. The results showed that the adoption of artificial intelligence significantly improved Green FinTech performance. The findings showed that sustainable innovation acted as a mediating factor, enabling firms to transform artificial intelligence capabilities into more efficient, environmentally sustainable, and cost-effective FinTech practices. These results indicated that energy sector firms that adopted artificial intelligence and invested in sustainable innovation were better positioned to improve Green FinTech performance and achieve long-term competitive advantages. The study offered important implications for managers, investors, and decision-makers seeking to accelerate the digital and sustainable transformation of the energy sector.