Data Intelligence and Informatics Practice

Industrial Artificial Intelligence: A Review of Technologies, Applications, Challenges, and Future Directions

Rajiv Kumar Verma — 2026-08-04

Industrial Artificial Intelligence (IAI) has emerged as a transformative paradigm that integrates advanced computational intelligence with industrial systems, enabling unprecedented levels of automation, optimization, and decision support across manufacturing and process industries. This review provides a comprehensive synthesis of IAI technologies, spanning machine learning, deep learning, computer vision, natural language processing, reinforcement learning, digital twins, and explainable AI, and examines their deployment across diverse industrial sectors including automotive, aerospace, energy, logistics, and semiconductor manufacturing. Key application domains, including predictive maintenance, quality control, process optimization, and supply chain management, are critically analyzed in terms of methodology, performance benchmarks, and deployment maturity. Unlike prior reviews that focus primarily on a single technology family or a single application domain, this review integrates technology, application, and challenge perspectives within one evidence-based framework and benchmarks technology maturity against deployment readiness across eight industrial sectors. The review further identifies persistent challenges such as data scarcity, black-box opacity, legacy system integration, cybersecurity vulnerabilities, and workforce readiness, and proposes evidence-based mitigation strategies for each. A forward-looking perspective is offered, highlighting the convergence of large language models, federated learning, and Industry 5.0 human-centric frameworks as defining trajectories for the next decade. Drawing on 40 peer-reviewed journal articles identified through a systematic literature search, this article serves as a structured foundation for researchers, engineers, and policymakers navigating the complex IAI landscape.

Comparative Analysis of Machine Learning Models for Credit Risk Assessment in FinTech Applications

Ayesha Tariq — 2026-08-05

Credit risk assessment is a core function of financial technology (FinTech) platforms, underpinning lending decisions, pricing, and portfolio risk management. Over the past decade, machine learning (ML) has progressively displaced traditional statistical scorecards as the dominant analytical paradigm, driven by the availability of large, high-dimensional, and often alternative data sources together with the need for higher predictive accuracy. This review provides a comparative analysis of ML models used for credit risk assessment in FinTech applications, spanning traditional statistical classifiers, tree-based ensemble methods, and deep learning architectures. Following a structured literature search and screening process, we synthesize empirical evidence on predictive performance across benchmark and proprietary FinTech datasets, critically contrast model families under varying data conditions, and examine the trade-off between predictive power and interpretability that shapes real-world adoption. We further discuss the role of alternative data and explainable artificial intelligence (XAI) in expanding financial inclusion while preserving regulatory compliance, distinguishing global from local interpretability, and we outline open challenges, including data imbalance, concept drift, algorithmic fairness, privacy, and model governance. The review concludes with directions for future research, emphasizing hybrid and federated learning architectures, standardized benchmarking protocols, and explainability-by-design frameworks for credit risk applications in emerging FinTech markets.

Decision Support Systems: Evolution, Architectures, Applications, and Future Directions

Emeka J. Eze — 2026-08-04

Decision Support Systems (DSS) have evolved from stand-alone model-based tools into integrated, data-intensive, and increasingly intelligent systems that assist decision makers in healthcare, finance, supply chains, public administration, and other complex settings. This review article presents a structured integrative synthesis of foundational and contemporary DSS literature. The review examines conceptual origins, major taxonomies, architectural components, domain applications, implementation constraints, and emerging research directions. The synthesis shows that modern DSS combine data management, analytical models, knowledge representation, collaborative functions, and artificial intelligence, but their effectiveness remains contingent on data quality, organisational fit, explainability, interoperability, cybersecurity, and meaningful human oversight. Rather than comparing heterogeneous studies through unsupported aggregate performance statistics, this review evaluates evidence according to study context, DSS function, reported outcome, and methodological limitations. A divergent-convergent process model is presented to explain how multiple analytical perspectives can be generated, integrated, assessed, implemented, and refined through feedback. The review identifies priority directions in real-time adaptive support, interoperable architectures, explainable and accountable AI, longitudinal human-factor evaluation, and governance of increasingly autonomous decision systems. The article contributes an integrated framework that connects DSS evolution, architecture, organisational conditions, risks, and future development.

Big Data Analytics for Customer Behavior Prediction in E-Commerce Platforms in Kenya

Peter Kiprono — 2026-08-05

Big data analytics (BDA) has emerged as a transformative tool for decoding complex consumer behaviour patterns in digital marketplaces. However, its application within the context of emerging African economies, particularly Kenya, remains underexplored despite the rapid expansion of mobile-enabled e-commerce platforms. This study investigated the efficacy of BDA techniques in predicting customer purchase behaviour across Kenyan e-commerce platforms, combining a structured survey of 400 registered online shoppers with transaction log data extracted from three major platforms: Jumia Kenya, Kilimall, and Masoko. Five machine learning algorithms (Logistic Regression, Decision Tree, Random Forest, XGBoost, and Long Short-Term Memory, LSTM) were trained and evaluated using stratified 10-fold cross-validation and a held-out test partition, with a Hybrid Random Forest + XGBoost ensemble achieving the highest test-set classification accuracy of 93.1% (95% CI [83.8%, 97.2%]; F1-score = 92.8%). Key predictors identified through SHAP-based feature importance analysis included purchase frequency, session duration, cart abandonment rate, and M-Pesa transaction size, reflecting the mobile-payment-centric nature of Kenyan consumer behaviour. Platform-level analysis revealed that M-Pesa facilitated over 70% of transactions, and sentiment analysis of customer reviews yielded a positive polarity rate of 62.3% (95% CI [59.5%, 65.0%]). The findings demonstrate that culturally and infrastructurally contextualised BDA models can achieve accuracy levels competitive with those reported in technologically mature markets and provide actionable insights for platform optimisation, personalised recommendation delivery, and customer retention strategy in sub-Saharan digital markets.

Adoption and Sectoral Impact of Data Science Practices in India: A Cross-Sectional Case Study

Karthik Subramanian — 2026-08-04

Data science has become a defining capability for organisations seeking to convert large volumes of structured and unstructured information into actionable decisions, yet the extent and pattern of its adoption across India's diverse industrial landscape remain incompletely documented. This study reports a cross-sectional survey of 420 organisational respondents drawn from seven sectors—information technology and IT-enabled services (IT/ITES), banking, financial services and insurance (BFSI), healthcare, government and public sector, agriculture and allied industries, retail/e-commerce, and education—across metro and non-metro locations in India. A structured, validated questionnaire was used to capture adoption levels, tools and techniques in use, perceived barriers, and organisational impact. Results showed that 55.0% of organisations had reached at least an intermediate level of data-science adoption, with marked sectoral disparity: IT/ITES and BFSI organisations reported the highest adoption (81.3% and 72.6% intermediate-to-advanced, respectively), while agriculture and allied industries and education lagged considerably (21.3% and 23.4%). Python, SQL and spreadsheet tools dominated the technology stack, while distributed big-data frameworks and natural-language-processing tools remained comparatively underused. Shortage of skilled manpower (74.3%), data-privacy concerns (65.2%) and poor data quality (52.6%) were the most frequently cited barriers. Adopting organisations reported the strongest perceived gains in decision-making (mean = 4.21/5) and operational efficiency (mean = 4.05/5). A five-year retrospective trend indicated accelerating adoption across all sectors between 2020 and 2025, with the gap between leading and lagging sectors widening rather than narrowing. These findings suggest that targeted workforce development, clearer data-governance frameworks and infrastructure investment in lagging sectors are needed to translate India's data-science momentum into more broadly distributed organisational value.